Predicting Fall Risks in Electronic Patient Records Ariane Schenk, Sebastian Ahrndt, and Sahin Albayrak DAI-Labor, TU Berlin Ernst-Reuter-Platz 7 10587 Berlin [email protected] (Corresponding author) Abstract: The increasing progress in the development and deployment of electronic patient records (EPR), especially those for mobile devices like tablets, offers new opportunities for the health care sector. In the field of prevention, health care professionals can gain benefits from the consolidation of complementary data sources and the ubiquitous access to patient information. Hence, this work presents the results of a literature research done to identify fall indicators and their influence on patients fall risk. This work takes special interest in the specific needs of home visiting nurses and the data available through a mobile EPR.

1 Introduction The prevention of falls is an important task in the health care of elderly people. Injuries or disabilities contracted by falls constitute not only as a cause of suffering of the elderlies, but also lead to substantial economic burden which evolve directly or indirectly from falls. For example, in 2005 the health care costs for fall related hip fractures led to direct expenses of 2.77 billion Euro in Germany [WG07]. Due to the demographic change these costs are projected to increase to 3.85 billion Euro a year in 2030. However, the increasing development of electronic patient records (EPR), especially those for mobile devices like tablets, offers new opportunities not only in the documentation, but also in the prevention of falls. Hence, in this work two questions are discussed: Firstly, which fall risk indicators are well-researched and allow a prediction of the fall risk of community-dwelling elderly and secondly which of them are already available through the dataset of contemporary EPR in ambulant heath care? We integrated the results of this work into the agneszwei App, an EPR developed especially for the requirements of home visiting nurses [ARA12]. Although the integration process is not within the scope of this work, we will present the already implemented connections between the identified fall risk indicators and the dataset available through the used EPR (See Section 3). Afterwards, we will conclude this work and present an outlook to the future work (See Section 4).

1194

2 Methodology In order to identify fall risk indicators, we carried out a literature research in March and April 2012 using the keywords fall“, risk factors“, older people“, community” ” ” ” dwelling“ and their German equivalents. The databases PubMed, DIMDI and GoogleScholar were queried, complemented by research within relevant literature. Further interesting publications were found by specifically looking into the lists of references. Firsthand studies, meta-analysis and guidelines from Europe and North America were considered. Only publications covering a community-dwelling setting were taken into account or publications clearly pointing out the differences between institutionalized and ambulant settings.

3 Results Given the breadth of research results in the field of fall risk and fall prevention, we started our study with a requirement analysis interviewing both developers and members of the potential target audience, who are home visiting nurses. Furthermore, we gained some insights into the application itself and the available data provided by the EPR (for detailed informations about the EPR, see e.g. [ARA12]). We identified several requirements, which revealed usability to be one of the most important. To ensure acceptance by the potential target audience, the application should help facilitate the work and ease the nurse’s work. To fulfill this requirement we decided to find the major categories of fall risk indicators which therefore cover the majority of all fall reasons. Following Robbins et al. [RRJ+ 89], we argue that it is sufficient to cover the major categories to calculate an adequate fall risk. Furthermore, there are at least two advantages: The major categories are well-researched and we prevent additional administrative burden on the nurses. Table 1 illustrates the identified categories and an related subset of considered studies1 . We were able to identify six different categories and 25 subcategories. The first category are physical deficits. They reach from weakness and visual, walk and balance deficits to the increase of fall risk upon self-perceived mobility limitations [GBSR07, RRJ+ 89]. Following Anders et al. [ABDW08], we argue that disorders of the locomotor system are the most important fall risk indicator. Secondly the category of cognitive deficits can be divided into fall fear as well as cognitive impairments and (temporary) dizziness. The third category are drugs, which is divided into two kinds of drug types which strongly influence the fall risk and a third which indicates the consumption of more than four different types of drugs. One may argue, that for example antiepileptics, antiarrhythmics or diuretics are missing as important subcategory here. However, we identified several studies discussing whether they are or they are not an important indicator (see, for example [TSD+ 98] and [WBB11]). Due to these controversial results, some future research in this area is needed. Diseases are divided into five subcategories, classifying the most fall influencing diseases. 1 Throughout the limited space, we decided to shorten the reference list. Please follow this link for the full list of references: http://goo.gl/vwalQ

1195

Tabelle 1: Identified categories and subcategories of fall risk indicators and a subset of references (sorted in descending order by year) accomplished by the number of studies related to each subcategory (out of 18 considered studies). The collected data were clustered throughout their prevalence (quantification and operationalizability) and their influence on the fall risk.

Category

Subcategory

References

Deficit (Physical)

Weakness (Extremities) Weakness (Common) Visual Deficit Walk Deficit Balance Deficit Mobility Limitation

[BR11, GBSR07] [BR11, Rub06] [BBSL12, BR11] [BBSL12, BR11] [BBSL12, GBSR07] [GBSR07, Rub06]

6 4 11 11 11 6

Mental State and Functional Disability

Fear Cognitive Impairment Vertigo / Dizziness

[BBSL12, Rub06] [BBSL12, BR11] [RJ06, HRD04]

4 13 4

Drugs

Hypnotics & Sedatives Psychotropic More than 4 different

[BBSL12, BR11] [BBSL12, GBSR07] [BBSL12, GBSR07]

5 7 4

Disease

Dementia Depression Alcohol Dependence Joint Disease Incontinence

[GBSR07, MSP+ 06] [MSP+ 06, RJ06] [MSP+ 06, FCC+ 04] [BBSL12, RJ06] [BBSL12, BR11]

3 6 3 4 6

Extrinsic Factors

Clothing Footwear Habitation

[BBSL12, MSP+ 06] [BBSL12, MSP+ 06] [BBSL12, MSP+ 06]

3 4 4

Impaired ADL

12 ADL types (e.g. [SSU09])

[GBSR07, MSP+ 06]

10

Others

Age Sex Weight Earlier Falls

[BR11, GBSR07] [RBS+ 02] [FCC+ 04, RBS+ 02] [BBSL12, BR11]

6 1 2 12

1196

#

Following Richter et al. [RBS+ 02] and Rubenstein and Josephson [RJ06] there are several other diseases influencing the fall risk for example Parkinson and Stroke [MSP+ 06]. However, more than 90% of falls are indicated by multiple factors [BS98] and we are able to cover the majority of diseases using other categories and subcategories (e.g. Parkinson which causes (among others) walk and balance deficits in its course). Further, we found controversial results discussing multi-morbid illness (factor three and greater) (see, for example [BBSL12]). All previously introduced categories were intrinsic factors. Nevertheless, up to 20% of all falls are purely caused by extrinsic factors [BS98]. Therefore, the fifth category describes extrinsic factors summarizing indicators like ill-fitting clothing and footwear, bad lightning, trip hazards and other domestic equipment influencing the fall risk. The impairment of the activities of daily living (ADL) (see, e.g. [SSU09]) is the fifth category and the first one which will be completed by the application itself. This is done using a semi-automatic approach, as not all necessary information may be available through the EPR. In contrast, the sixth category –others– will be completed automatically as all necessary data is provided by the EPR.

4 Conclusion In this paper, we introduced indicators for the fall risk of community-dwelling elderly. To identify these indicators, we performed a literature research with respect to the specific requirements and the available data provided by an EPR developed for ambulant health care. As usability was revealed as one of the most important requirements, the results ensure that the application should help facilitate the work and disburden the home visiting nurse’s work. Hence, we were able to identify physical and cognitive deficits as the most important categories as they are influenced by the other categories. Nevertheless, as the majority of falls are indicated by multiple factors, the other categories are just as important as fall risk indicators. To facilitate the work of the nurses we started to implement semi-automatic and automatic services, helping to estimate a patients fall risk. These services are based on the data available through the EPR. In the future, we will extend these capabilities using not yet considered data sources (for example symptoms of diseases or adverse effects of drugs). Also, the literature research results will be used to specify an metric which should ease the judgment of the fall risk. In order to prove this judgment future work will discuss the question whether a valid fall risk predicting instrument helps to prevent falls or not?

Literatur [ABDW08] J. Anders, M. Behmann, U. Dapp und U. Walter. St¨urze im Alter. In Beweglich?, Jgg. 2007/20 of Weißbuch Pr¨avention, Seiten 167–181. Springer Berlin Heidelberg, 2008. [ARA12]

S. Ahrndt, A. Rieger und S. Albayrak. Entwicklung einer mobilen elektronischen Pa-

1197

tientenakte f¨ur die ambulante Versorgung in l¨andlichen Regionen. In Mobile Informationstechnologie in der Medizin (Mocomed 2012), Seiten 1–15. Mocomed 2012, Sep 2012. Submitted to. [BBSL12]

K. Balzer, M. Bremer, S. Schramm und D. L¨uhmann. Sturzprophylaxe bei a¨ lteren Menschen in ihrer pers¨onlichen Wohnumgebung. Deutschen Institut f¨ur Medizinische Dokumentation und Informatin (DIMDI), 2012.

[BR11]

C. Becker und K. Rapp. Falling in Geriatrics: Diagnosis and Treatment. Der Internist, 52(8):939–945, August 2011.

[BS98]

C. Becker und S. Scheible. St¨urze und sturzbedingte Verletzungen a¨ lterer Menschen. Fortschritte der Medizin, 116(32):22–29, 1998.

[FCC+ 04]

G. Feder, M. Clark, J. Close, C. Cryer, C. Czoski-Murray, D. Green, S. Illiffe, R.A. Kenny, C. McCabe, E. Mitchell, S. Mitchell, P. Overstall, M. Preddy, C. Swift und D. Wild. Clinical practice guideline for the assessment and prevention of falls in older people. Royal College of Nursing, November 2004.

[GBSR07] D.A. Ganz, Y. Bao, P.G. Shekelle und L.Z. Rubenstein. Will my patient fall? The Journal of the American Medical Association, 297(1):77–86, January 2007. [HRD04]

C. Heinze, U. Rissmann und T. Dassen. St¨urze bei a¨ lteren Menschen. Pflegewissenschaften, 2(2):105–110, 2004.

[MSP+ 06] M. Moers, D. Schiemann, P.Blumenberg, J. Schemann und H. Stehling, Hrsg. Expertenstandard Sturzprophylaxe in der Pflege. Expertenstandards. Deutsches Netzwerk f¨ur Qualit¨atsicherung in der Pflege, 1. Auflage, February 2006. [RBS+ 02]

M. Richter, C. Becker, J. Seifert, F. Gebhard, O. Pieske, M. Holch und G. Lob. Pr¨avention von Verletzungen im Alter. Der Unfallchirurg, 105(12):1067–1087, Dezember 2002.

[RJ06]

L.Z. Rubenstein und K.R. Jospehson. Falls and Their Prevention in Elerly People: What Does the Evidence Show? The Medical Clinics of North America, 90(5):807– 824, September 2006.

[RRJ+ 89]

A.S. Robbins, L.Z. Rubenstein, K.R. Josephson, B.L. Schulmann, D. Osterweil und G. Fine. Predictors of Falls Among Elderly People: Results of Two Population-Based Studies. Arch Intern Med., 149:1628–1633, March 1989.

[Rub06]

L.Z. Rubenstein. Falls in older people: epidemiology, risk factors and strategies for prevention. Age and Aging, 35(2):37–41, September 2006.

[SSU09]

S. Schewior-Popp, F. Sitzmann und L. Ullrich, Hrsg. Thiemes Pflege: Das Lehrbuch f¨ur Pflegende in Ausbildung. Thieme Stuttgart, 11. Auflage, M¨arz 2009.

[TSD+ 98]

A.M. Tromp, J.H. Smit, D.J. Deeg, L.M. Bouter und P. Lips. Predictors for falls and fractures in the longitudinal aging study Amsterdam. Journal of Bone and Mineral Research, 13(12):1932–1999, December 1998.

[WBB11]

Martin Wehling, Heinrich Burkhardt und Heinrich Burkhardt. Pharmakotherapie und ¨ geriatrische Syndrome. In Arzneitherapie f¨ur Altere, Seiten 219–254. Springer Berlin Heidelberg, 2011.

[WG07]

E.J. Weyler und A. Gandjour. Socioeconomic Burden of Hip Fractures in Germany. Gesundheitswesen, 69(11):601–606, 2007.

1198