A Case Control Study of
Pediatric Falls
U
sing EMR
Patricia R. Messmer, PhD, RN-BC, FAAN
Consultant- Nursing Education & Research
Broward Health- Broward General Medical Center
Chris Evert Children’s Hospital, Fort Lauderdale, Fl
Miami Dade College, Miami, F
L
Arthur R. Williams, PhD, MA, MPA
College of Public Health
University of South Florida, Tampa, FL
Michele Fix, RN, BSN
Children’s Mercy Hospitals and Clinics
Kansas City, MO
• The Joint Commission Safety Goal 9B
– Organizations must have a fall reduction program
– Interventions must be implemented to reduce patients’ fall risk factors
– Fall reduction strategies must be individualized
– Education & training regarding fall reduction programs must include
• Staff
• Patients/Families
– Program must be evaluated to determine effectiveness
• Evaluation of the patient population, settings & services
• Institute for Healthcare Improvement (IHI) 5 million Lives
Campaign
– Validated the need to decrease the number of hospital medical injuries
• Falls
• Patient safety across the lifespan
Child Health Corporation of America (CHCA)
Pediatric Falls Study
Patricia Jamerson, PhD, RN (PI)
• 28 children’s hospitals
• 782 pediatric patients fell
during a 6 month period
– slipped, tripped and fell from
their cribs, beds, chairs and
examination tables…
• 10 (1.3%) were dropped by
Falls Definitions
• Patient Falls-An unplanned descent to the floor, either with
or without injury to the patient (ANCC Magnet Manual
2008).
– Adjusted Falls-An unintended event resulting in a person
coming to rest on floor or other lower level (witnessed)
or reported to have landed on the floor (un-witnessed)
• Falls Rate - total # of patient falls x 1,000 divided by total
# of patient days (ANCC Magnet Manual 2008).
• Children At-High Risk For Falls
– Preschoolers
– Children > 10 are twice as likely to fall compared with
the total population; age group is related to serious
injury and death
– Children with disabilities and minimal mobility
– Children in wheelchairs, regardless of cognitive ability,
due to tips and falls
King’s Theory of Goal Attainment
• GROWTH AND DEVELOPMENT
– Pediatric falls differ according to stages of growth and development. • PERCEPTION
– A process of organizing, interpreting and transforming information from sense data and memory; behavior flows from one’s perceptions and influences one’s behavior.
• COMMUNICATION
– Communication between nurse and pediatric patient and their family • INTERACTION
– A process of perception and communication between person and environment, between person-to-person, represented by verbal and nonverbal behaviors that are goal directed
• TRANSACTION
– A process of interaction in which human beings communicate with environment to achieve “Mutual Goal Setting”
This Study
Purpose
– This study was part of a CHCA multi-site study in
which 31 children who fell during a 6 month period
in 2008 were submitted for data analysis
– Control subjects were not part of the CHCA study
– Reviewed cases for the year 2008 in this study
• 74 cases
• 242 inpatient randomly selected control cases
Methods Used in This Study
•
Research design: a retrospective observational case-control study
• Occurrence reports
• Electronic medical patient records used
• Pediatric fall cases differ from adult cases
Developmental stages are germane in analyzing falls*
• i
nfant (<1)
• toddler/pre-school (1-3)
• early childhood (3-6)
• middle childhood (6-12)
• teenagers (13-18).
Hockenberry, M. & Wilson, D (eds) (2009) Wong’s essentials of pediatric nursing 8th ed; St Louis: Elsevier
Research Questions
1. Do children differ in their “risk” of a fall
according to HDFS risk scores in a
case-control study?
2. Do the cases/controls fall as predicted
consistent with HDFS risk scores in a
case-control study?
Developing a Pediatric Falls Scale
• Difficulties in developing a valid, easy to use and
reliable risk assessment instrument
– Falls are relatively infrequent events in pediatrics
– Falls are associated with developmental stages, i.e.
toddlers learning to walk and in a rush
– Falls may be related to psychosocial variables not
just the medical condition of the child
• Failure to properly identify real risks on an risk
scoring tool may arise due to effectiveness of existing
Fall Prevention programs.
• Fall data is collected retrospectively and depends
heavily upon accuracy of medical records
Falls Assessment Instruments
• Adult Tools
– Morse Fall Scale
– Hendrich
– Tinetti
• Pediatric Tools
– CHAMPS
– Graf-PIF Scale
– Hendrich II
– Humpty Dumpty Falls Scale (Miami Children’s Hospital)
– I’m Safe (The Children’s Hospital of Denver)
Elaine Graf (Graf-PIF Scale) Children’s
Memorial Medical Center
• Length of Stay
•
1-4 days- score 0
•
5-9 days- score 1
•
10-14 days- score 2
•
15-19 days- score 3
• Children without an IV- score 1
• PT/OT ordered- score 1
• If prescribed anti-Seizure Medication score 1
• Acute or chronic Orthopedic diagnosis- score 1
• Falls classification
•
accidental falls
•
unanticipated physiological falls (Morse, 1977)
•
anticipated physiological falls
Note: Reported sensitivity 75%; specificity 76% at a cut-point of 2, but not documented
Humpty Dumpty English nursery rhyme
Humpty Dumpty sat on a wall, Humpty Dumpty had a great fall
All the king’s horses and all the king’s men
Couldn't put Humpty Dumpty together again
Humpty Dumpty Falls Scale (HDFS)
• Age
• Gender
• Diagnosis
• Cognitive impairments
• Environmental Factors
• Response to Surgery /
Sedation/Anesthesia
• Medication usage
Components Criteria Score (circle)
<3 4 3-7 3 7-10 2 >13 1 Male 2 Female 1 Neurological Diagnosis 4 Alterations in Oxygenation 3 Psych / Behavioral 2 Other Diagnosis 1 Not aware of limitations 3 Forgets limitations 2 Oriented to own ability 1 History of Falls or Infant-Toddler
placed in bed 4
Pt uses assistive devices or
Infant-toddler in crib 3 Patient placed in bed 2 Outpatient area 1 Within 24 hours 3 Within 48 hours 2 More than 48 hours/None 1 Multiple usage of:
Sedatives (excluding ICU patients sedated and paralyzed) Hypnotics Barbiturates Phenothizines Antidepressants Laxatives/Diuretics Narcotics
One of the meds listed above 2 Other medications/None 1 Total Score 3 Medication Usage Response to Surgery / Sedation Anesthesia Age Gender Diagnosis Cognitive Impairments Environmental Factors
A patient is considered at
risk for falls if score =>12
Range- 7-23
Maximum Score 23;
Minimum Score 7
Results: Demographics
Group Developmental
Level # Gender # Ethnicity Diagnosis
Pediatric patients who fell (74) CASES >3 infants 3-7 toddlers /pre-school 7-13 school aged 13+ teenagers 21 18 13 22 female Male 42 32 Predominately Caucasians Less than 33% Blacks, Hispanic or Asian #1 Gastrointestinal #2 Neurologic/Development Delay #3 Oncology #4 Orthopedic Pediatric patients who did not fall (242) CONTROLS >3 infants 3-7 toddlers /pre-school 7-13 school aged 13+ teenagers 80 44 45 73 female Male 65 83 Predominately Caucasians Less than 35% Blacks, Hispanic or Asian Other (middle eastern families do not mark themselves in these categories #1 Gastrointestinal #2 Oncology #3 Neurologic/Developmenta l Delay
Results
• All pediatric patients were assessed for falls
• Some actual falls not rated as High Risk using HDFS >=12
• 65% of control cases identified as at High Risk >=12.
• The OR was 1.15 with CI; .39, 3.15, p >.76.
• At a cut-off point of >=12 (High Risk per HDFS), sensitivity was
57%; specificity 39%.
• At a cut-off point of >=11 (not High Risk per present HDFS
scoring), sensitivity was 75%; specificity 28%.
A ROC CURVE ILLUSTRATING TRADE-OFFS
BETWEEN
SENSITIVITY AND SPECIFICTY
Miami Children's data
Legend; The top line is a hypothetical curve where sensitivity (vertical axis) = .80 and specificity (horizontal axis) = .80 at the inflection point. The solid line with the dots is the empirically determined HDFS ROC which equals only 54% of the area of the rectangle of this ROC curve.
HDFS Results
• The correlation between HDFS and age is r = - 0.52 (p<.001)
• An OLS regression indicated that age is highly associated with lower
HDFS.
• Age alone in this regression accounts for 32% of the variance in HDFS
• These results suggest that age should be better accounted for than it
is at present in the HDFS.
• However, the ages or age groups should be assigned empirically
determined weights, e.g, a strictly linear relationship (straight line) may
not identify risks of falling as well as a step function shown below.
• Rescoring of the HDFS with empirically determined weights of all risk
items may improve results.
Ordinary Least Squares Regression:
Age on HDFS scores
5 10 15 20 25 0 5 10 15 20 age 95% CI Fitted values hdfsLegend: R
2=0.346, F= 154.30, p<.001; coefficient for age in
Contributors to Pediatric Patient Falls
and Fall-Related Injuries
(Woods, et al., 2005)
Preventable adverse events:
falls, near falls, fall Injuries Risk of a Fall
Child’s Human factors Nurse’s human factors Latent system factors Biomechanical factors Fall Environment factors
Parent’s human factors
Pediatric Fall Harm Index (PhFX) assesses the collective
harm occurring from falls in an objective manner
Weight Level of
Harm
Clinical Criteria
0
No Harm
No Assessment, diagnostic
testing or treatment required
1
Minimal
Harm
Superficial assessment &/or
treatment required for injury (i.e.
cleaning of site, ice , bandage)
2
Moderate
Harm
Assessment, diagnostic testing,
&/or treatment required for injury
(i.e. laceration or fracture
suspected or diagnosed)
3
Significant
Harm
Assessment, diagnostic testing,
&/or treatment required for
permanent harm (i.e. brain/spinal
injury, or death)
Minor injuries (19%) in 74 falls
Fall Injury Rate
Minor Harm, 19%
No Harm, 81% Death, 0%
Major Harm, 0% Moderate Harm,
0%
Discussion
• The HDFS captures some of the real risk of falling among
hospitalized pediatric patients as shown by the odds ratio
(OR) at a cut point of 12.
• However, further assessment of the instrument is needed
for efficient screening.
• Far too many false positives are identified as indicated by
the poor specificity of the instrument.
• The smaller number of false negatives also requires some
“fix” to raise sensitivity about or above .80.
• Nurses must monitor pediatric patients frequently, record
HDFS scores in EMR, for assessment, and implement
preventive fall measures.
Discussion
• Nurses may not always observe their patients holistically but
rely on the patient’s present condition.
• Nurses may not take into account underlying factors that place
patients at higher risk for falls such as behavior.
• The HDFS requires further testing and development to be a
reliable screening tool, but this seems to be the case for other
tools now available.
• At a minimum, weights of risk “parameters” should be
empirically determined and more parameters may be needed
on tools, particularly behavioral parameters.
• The HDFS is a promising tool for research and further
development is needed in clinical settings, but the tool might
be used in clinical settings
i
f leaders and staff nurses are
cognizant of its potential for over-identification of the risk
status of patients.
Conclusion
• A nurse’s best clinical judgment remains a valuable resource in decreasing the incidence of falls and related injury.
• Nurses and parents must increase their awareness of patient injuries from falls in order to provide safe, non-invasive care.
• Preventing falls is challenging due to pediatric patients’ cognition, growth, and development.
• Although pediatric fall rates are well below adult fall rates, falls may not be as carefully monitored in pediatric as compared to adult
facilities.
• Fall rates are derived from voluntary reporting mechanisms and may vary due to reporting rather than the actual fall rate number per
1000 patient days.
• Preventing falls in pediatric populations is challenging due to the
unpredictability of falls related to a child’s age, cognition, growth, and development.
Lessons Learned
• Using the HDFS fall tool as part of pediatric assessment
stresses the need for nurses’ best clinical judgment, a
valuable resource in decreasing fall incidence/related injury.
• Identifying patients at-risk for falls ensures that all
disciplines, parents and visitors increase their awareness of
patient injuries in providing safe, noninvasive care.
• Interventions should be based on developmental levels.
• Enhance vigilance for all pediatric patients in room, hallway
and playroom.
• Observations
• Parent-child interaction
• Parenting skills
Recommendations for Further Study
• Future studies with larger sample sizes across multiple institutions identifying children at risk on HDFS, focusing on preventing serious injury
• Attention needs to be given to the weights assigned to “parameters” on the HDFS and the possible incorporation of
– parental presence or even parental “risks” – behavioral characteristics of the patient. • Improve the scoring algorithm
– Use a greater number of potentially predictive variables
– Use the HDFS prospectively with patients to determine clinically relevant screening properties.
• Why would we (pediatric nurses) expect parents/guardians to heed our words on children’s safety while in their care, if we can not keep them safe while in our care?” Decreasing number of pediatric falls and
implementing falls prevention programs is every pediatric nurse’s responsibility
Review of Literature
• Child Health Corporation of America (2009). Pediatric falls: State of the science. Pediatric Nursing 35(4), 227-231.
• Frankenburg, W.K., Camp, B.W. (1975). Pediatric screening tests. Springfield, IL: Thomas.
• Graf, E. (2011). Magnet Children’s Hospitals: leading knowledge development and quality standards for inpatient pediatric fall prevention programs. Journal of Pediatric Nursing 26(2), 4-5.
• Haynes, R.B., Sackett, D.L., Guyatt, G.H., Tugwell, P. (2006). Clinical epidemiology: how to do clinical practice research. 3rd ed; Boston, MA: Lippincott, Williams & Wilkins.
• Hendrich, A’, Bender, P, Nyhuis, A. (2003). Validation of Hendrich II Fall Risk Model: A large concurrent case/control study of hospitalized patients. Applied Nursing Research, 16(1), 9-21.
• Hockenberry, M. & Wilson, D. (eds) (2009) Wong’s essentials of pediatric nursing 8th ed. St Louis. Elsevier.
• Hill-Rodriguez, D., Messmer, P.R., Williams, P.D., Zeller, R.A.,
Williams, A.R., Wood, M. & Henry, M/ (2009). Humpty Dumpty Falls Scale: A case control study. JSPN 14(1), 22-32.
• Kingston, F., Bryant, T., & Speer, K. (2010). Pediatric falls benchmarking collaborative JONA 40(6), 287-292.
Review of Literature
• Meyers, H.& Nikoletti, S. (2003). Fall risk assessment: A
prospective investigation of nurses’ clinical judgment and risk assessment tools in predicting patient falls. IJNP, 9, 158-165.
• Monson S. et al (2008). In-hospital falls of newborn infants: data from multihospital healthcare system Pediatrics 122(2) e277-e280. • Morse, J.M. (2006). The safety of safety research: the case of
patient fall research. Can J Nurs Res 38(2), 73-88
• Neiman, J. & Rannie, M. I’m safe: development of a fall prevention program to enhance quality and patient safety. 8th NICHQ
conference, Grapevine, TX, March 11, 2009.
• O’Connell, B., Myers, H. (2002). Research in brief. The sensitivity and specificity of Morse Fall Scale in an acute care setting. Journal of Clinical Nursing, 11(1), 134-135.
• Pillai, S.B.(2000). Fall injuries in pediatric population: Safer and most cost effective management. Journal Trauma, 48(6),1050-51. • Raszmus, I. (2006). Falls in hospitalized Children. Pediatric Nursing
32(6), 568-572.
• Robles, J. (2009). The effect of the electronic medical record. Creative nursing. 15(1), 31-35.
• Woods et al. (2005). Anatomy of a patient safety event: a pediatric safety taxonomy Qual Saf Health Care 15, 422-427