Research Article Appraisal

REVIEW PAPER

Meteorological factors to fall: a systematic review

K. P. Chow1 & D. Y. T. Fong1 & M. P. Wang1 & J. Y. H. Wong1 & Pui Hing Chau1

Received: 3 July 2017 /Revised: 10 September 2018 /Accepted: 2 October 2018 /Published online: 27 October 2018 # ISB 2018

Abstract There existed systematic review on studies investigating the association between hip fractures and external risk factors including meteorological factors. Albeit the fact that most serious common fall injury is a hip fracture, it cannot account for all injuries forms of fall. There was a lack of systematic review covering all fall-related injury or deaths to thoroughly summarise meteorological aspects of fall. This study aimed to systematically review epidemi- ological studies of fall and fall-related circumstances without restriction to hip fracture. A systematic search in three databases, namely PubMed, CINAHL Plus and EMBASE, was performed. Searches in two Chinese databases named the Wanfang Med Online and the China Journal Net were done in addition. A total of 29 studies were identified. The study site, fall cases identification, meteorological factors and findings of all the selected studies were being extracted. The quality of the studies was critically appraised. We identified some of the environmental risk factors to fall among those studies. Ranging from the lower ambient temperature, the presence of snow cover, seasonal factors, and time of the day to location of fall, these factors have different levels of impact related to higher incidence or mortality of fall. To conclude, a better understanding of injury mechanisms is a prerequisite for preventive interventions.

Keywords Fall . Meteorological factors . Temperature . Weather

Introduction

Fall is Ban event during which a person takes a lying position on a lower or the same level without the loss of consciousness^ (National Insitute for Health and Care Excellence 2004). Fall constitutes a crucial element of injury or death particularly among the older ages (Nagurney et al. 1998; Sterling et al. 2001; Zautcke et al. 2002). Most of the injuries among older people are caused by falls in northern countries like Finland and Norway (Bulajic-Kopjar 2000; National Center for Injury Prevention and Control 2004; Saari et al. 2007). In northern countries like Canada and England, there were 27% of community-dwelling persons over age 65 years who fell each year and the figure rose to 50% for those over 80 years old (Carpenter 2009; Stalenhoef et al. 1997). Among the community-dwelling older people in United States, almost 30–40% of those aged over 65 years fall

at least once in a year (Mertz et al. 2010; Rubenstein and Josephson 2002). Also, in an Asian city of Hong Kong which has an ageing population, the prevalence having at least one fall in the preceding 12 months among the elderly was be- tween 18 and 19.3% (Chu et al. 2005, 2007).

About 5–11% of falls lead to a serious injury requiring medical care and about 5% of falls results in a fracture (Kannus et al. 1999; Lilley et al. 1995). It is reported that unintentional falls are the leading cause of injury mortality and morbidity among adults aged 65 years and above in United States (National Center for Injury Prevention and Control 2004). Take China as an example of Asian countries, fall injury is the fourth fatal injury among all ages and is listed the top of fatal injury in the population aged 65 above (Liu et al. 2015). Even when a fall is not causing serious injuries, it may lead to anxiety, depression, loss of confidence and con- fusion (Kojima et al. 2008), affecting elders’ quality of life in a large extent. Epidemiological research on fall-related injuries is therefore essential for formulating effective prevention strat- egies that target specific risk factors (Mondor et al. 2015).

A fall is almost never caused by one factor, but is a conse- quence of the presence of several factors (S. R. Lord et al. 2001). Risk factors can be classified into intrinsic and extrinsic

* Pui Hing Chau [email protected]

1 School of Nursing, The University of Hong Kong, Pokfulam, Hong Kong

International Journal of Biometeorology (2018) 62:2073–2088 https://doi.org/10.1007/s00484-018-1627-y

(American Geriatrics Society, British Geriatrics Society, and American Academy of Orthopaedic Surgeons Panel on Falls Prevention 2001; Todd and Skelton 2004). Intrinsic risk fac- tors of fall are specific to individual which included biological factors like age and sex and behavioural factors such as med- ication use. Intrinsic risk factors of fall have been extensively studied and advanced age and female sex are two of the recognised ones (S. Lord et al. 2007). On the other hand, extrinsic factors are specific to socioeconomic factors like education and income level and environmental factors like flooring condition and lighting. Fall injuries can occur be- cause of a complex interplay of environmental factors (Bergstrom et al. 2008). The most common environmental causes are slippery floor and tripping over an obstacle (Ip and Ip 2006). Fall due to this mechanism is preventable by manipulating the environmental risk factors to decrease the occurrences of falls.

There are numerous investigations on the association be- tween hip fractures and its risk factors including the effect of seasonal changes. Most previous studies of seasonality fo- cused on hip fractures and results were inconsistent across all regions. Seasonal trends in fall or hip fracture hospitalisation counts have previously been reported in a number of countries, including the USA and Canada (Arbes and Berzlanovich 2015; Berg et al. 1997; Bulajic-Kopjar 2000; Campbell et al. 1988; Gevitz et al. 2017; Gyllencreutz et al. 2015; Hemenway and Colditz 1990; Jacobsen et al. 1995; Kojima et al. 2008; Lopez-Soto et al. 2016; Lund and Sheafor 1985; Luukinen et al. 1996; Magota et al. 2017; Modarres et al. 2012; Pipas et al. 2002; Ryynanen et al. 1991; Yeung et al. 2011). However, there are also several studies conducted in the United States and abroad reported with no seasonal variation in the incidence of fall or hip frac- tures (Bergstrom et al. 2008; Mondor et al. 2015; Saari et al. 2007; Stevens et al. 2007).

Albeit the fact that most common fall injury is a hip frac- ture, there are other injuries of fall or even death (Bulajic- Kopjar 2000). To thoroughly understand the aetiology of fall related to external factors, a complete systematic review is needed to identify meteorological factors of fall. Based on our findings, strategies for fall-related injuries involving envi- ronmental or behavioural modifications would be recom- mended. A systematic review revealed an association between hip fractures and temperature, snow, ice, and sun exposure (Roman Ortiz et al. 2015). Yet, this did not cover all fall cases, particularly the fatal ones. While there was review on climatic factors to hip fracture (Roman Ortiz et al. 2015), it did not focus on fall in general.

This study aimed to systematically review meteorological factors that focus on fall, with a view to identifying environ- mental factors of fall and informing policy makers, health care professionals and general public to adopt timely preventive measures in advance of a fall.

Methods

Search strategies

Three electronic databases including the PubMed, CINAHL Plus and EMBASE were used for this systematic review. Three groups of keywords were chosen to facilitate the searching process. The first one was Bfall^. The second group was Bincidence^, Bmortality ,̂ Bdeath^ or Bhospitalisation^. The third group was Bclimate^, Bweather^ or Btemperature^. Keywords within each group were linked together by OR, while the three groups were linked together by AND. Separately, we had done a search using the above keywords (in Chinese) in two Chinese databases named the Wanfang Med Online and the China Journal Net.

Inclusion criteria were (1) studies that mainly focus on fall accidents; (2) cohort studies; and (3) studies in English or Chinese. Studies about fractures or injuries without explicit relation to fall were excluded.

After initial screening and adding manual searches after screening the reference list of the selected articles, a list of the relevant studies was then used for the review. Duplicated studies were removed, followed by further screening of the title and abstract to identity relevant studies. The full text of all articles deemed to be relevant was then retrieved and screened according to the selection criteria. The reporting of the sys- tematic review followed the PRISMA checklist (Moher et al. 2009).

Data extraction and assessment of methodological quality

A table of evidence was used to summarise data related to study design, study quality, study site, study period, study population, fall case identification (inclusion and exclusion criteria), meteorological factors and result findings. A narra- tive interpretation of the results was chosen over a statistical analysis to carry out a comparison between the findings since there was some diversity in reporting of environmental factors and their statistical analyses.

Methodological quality was critically assessed accord- ing to SIGN checklist (Scottish Intercollegiate Guidelines Network 2015) for cohort studies. The overall methodo- logical quality of the study was rated after review. Review and the critical appraisal of the articles were car- ried out by the first author (KP Chow) and were further reviewed by the corresponding author (PH Chau). Both of them independently assessed the studies for inclusion in the review and performed the critical appraisal of the studies. All the decisions made were agreed by both re- viewers. Any disagreements were resolved by consensus through discussion.

2074 Int J Biometeorol (2018) 62:2073–2088

Results

Search history

The search was performed till March 2018. After the keyword search through the PubMed, CINAHL Plus and EMBASE, a total of 2015 studies were identified. No relevant studies were identified in the two Chinese databases. Forty-seven duplicat- ed copies were removed. Ten studies were added after manual search of reference lists. After screening of the titles and ab- stract, 42 of them remained. The full text of all articles deemed to be relevant was then retrieved. There were two cross- sectional studies excluded. Six studies were excluded because they were not in English or Chinese. Another five were ex- cluded as the studies did not specify whether the fractures/ injuries were due to fall. A total of 29 studies were included in the review. Figure 1 displays the flow diagram in accor- dance with PRISMA statement to show selection process (Moher et al. 2009).

Study quality assessment

Among the 29 studies, 2 of them got high quality (++). Remaining studies were classified as acceptable (+). All stud- ies were found to have met most of the criteria in the appraisal and would be included in the synthesis.

The main features of the selected studies were summarised in Table 1.

Publication year and study sites

The year of publication of the selected studies ranged from 1985 to 2017. The studies were mainly conducted in northern countries such as the United Kingdom (Parker and Martin 1994; Parker et al. 1996), Italy (Lopez-Soto et al. 2016), Germany (Arbes and Berzlanovich 2015), Finland (Luukinen et al. 1996; Saari et al. 2007), Sweden (Bergstrom et al. 2008; Gyllencreutz et al. 2015; Vikman et al. 2011), Norway (Bulajic-Kopjar 2000), Canada (Driedger et al. 2016; Mamdani and Upshur 2001; Modarres et al. 2012; Mondor et al. 2015; Morency et al. 2012) and the United States (Aharonoff et al. 1998; Centers for Disease Control and Prevention 2004; Gevitz et al. 2017; Hemenway and Colditz 1990; Jacobsen et al. 1995; Lund and Sheafor 1985; Pipas et al. 2002; Smith and Nelson 1998; Stevens et al. 2007). There were few studies from southern region. There was one study conducted in Taiwan (Lin et al. 2015), Hong Kong (Yeung et al. 2011), New Zealand (Campbell et al. 1988), Australia (Turner et al. 2011) and Japan (Magota et al. 2017), respectively.

Records identified through

database searching

(n = 2015)

Sc re en

in g

In cl ud

ed El ig ib ili ty

Id en

�fi ca �o

n Additional records identified through other sources

(n = 10)

Records after duplicates removed

(n = 1978)

Records screened

(n = 1978)

Records excluded after

screening titles and

abstract

(n =1936)

Full-text articles

assessed for eligibility

(n = 42)

Full-text articles

excluded

(n =13)

6 not in

English/Chinese, 5 not

specify fall as causes

and 2 not cohort studies

Studies included in the

qualitative synthesis

(n = 29)

Fig. 1 PRISMA flow diagram showing the selection process

Int J Biometeorol (2018) 62:2073–2088 2075

Ta b le 1

M et eo ro lo g ic al fa ct o rs an d fa ll : m ai n fe at u re s o f th e se le ct ed

st u d ie s

C it at io n /s tu d y

ty p e/ q u al it y

L o ca ti o n /p o pu la ti o n

O u tc o m e (s tu d y p er io d )

D ef in it io n

M et eo ro lo g ic al v ar ia b le s

R es u lt s

A h ar o n o ff et al .

(1 9 9 8 )/ co h o rt

st u d y /+

U .S ./ fa ll p at ie n ts

S u st ai n ed

a fe m o ra l ne ck

o r

in te rt ro ch an te ri c fr ac tu re

o f

n o n p at h o lo g ic o ri g in

(1 Ju l 1 9 8 7 to 3 1 D ec ,1 9 9 6 )

N .A .1

N .A .

M o st fr ac tu re s o cc u rr ed

at ho m e, p ar ti cu la rl y in

p at ie n ts w h o w er e o ld er , le ss

h ea lt h y, an d h ad

p o or er

am b u la to ry

fu n ct io n s (P

2 < 0 .0 0 1 )

M o re

th an

7 5 %

o f fr ac tu re s re su lt ed

fr o m

a fa ll

w h il e th e p at ie n t w as

st an d in g o r w al k in g

M o st fa ll s oc cu rr ed

d u ri ng

d ay li g h t h ou rs w it h a

p ea k in

th e af te rn o o n (3 8 .5 % ) (P

< 0 .0 0 1 )

N o se as o n al v ar ia ti o n

A rb es

an d

B er zl an o v ic h

(2 0 1 5 )

/C o h o rt st u d y /+

M u n ic h , G er m an y /f at al fa ll

p at ie n ts re q u ir in g au to p si es

F at al in ju ri es

as a re su lt o f fa ll s

F al li n g fr o m

o n e le v el to

an o th er

le v el o r o n th e

sa m e le v el

N .A .

O v er 5 5 %

o f ca se s fe ll fo rm

o n e le v el to th e sa m e

le ve l

M o re

m al e (5 9 % ) in v o lv ed

in fa ta l fa ll s th an

fe m al e (4 1 % )

61 %

ca se s w er e ag ed

6 5 +

F at al fa ll in ci d en ce

w as

h ig h er in su m m er m o n th s

(3 3 % ) th an

in w in te r m o n th s (2 4 % )

H ea d tr au m a w as

th e do m in an t ca se

o f fa ll d ea th

re g ar d le ss

o f th e h ei g ht

o f fa ll (4 7 % )

B er g st ro m

et al .

(2 0 0 8 )/ co h o rt

st u d y /+

U m eå , S w ed en /t ra u m a p a-

ti en ts ag ed

5 0 + y ea rs an d

o ld er

ad m it te d in

U m eå

U n iv er si ty

h o sp it al

em er g en cy

d ep ar tm

en t

F al l- re la te d fr ac tu re

(1 9 9 3 – 2 0 0 4 )

C o d es

o f th e E u ro p ea n H o m e

an d L ei su re

A cc id en t

S u rv ei ll an ce

S y st em

(E H L A S S )

N .A .

46 %

in d oo rs an d 4 4 %

o u td o o rs

T h e m o n th ly

ra te q ui te st ab le

L es s in fl u en ce

fr o m

cl im

at e (s no w , ic e) , h o u rs o f

d ay li gh t/ d ay

fo r o ld

ag e g ro u p

B u la ji c- K o p ja r

(2 0 0 0 )

/C o h o rt st u d y / + +

T hr ee

u rb an

ar ea s in

N or w ay

(S ta v an g er ,T

ro n d h ei m ,a n d

H ar st ad )/ p o p u la ti on

ag ed

6 5 + ye ar s an d o ld er

F al l- re la te d fr ac tu re s (1

Ja n 1 99 0 to

3 1 D ec

1 9 9 7 )

F o r n at u re

o f in ju ry , ca se s w er e

g ro u p ed

b as ed

o n IC D -9

E ig h t co ld er

se as o n s (f ro m

1 O ct o b er

th ro u g h 31

M ar ch ), an d d u ri n g th e

ei gh t m il d er

se as on s

(f ro m

1 A p ri l th ro u g h 30

S ep te m b er )

R R 3 fo r in ju ri es

(c o ld er

se as o n s v s m il d er

se as o n s) :

A g ed

6 5 – 7 9 = 1 .3 9 (9 5 %

C I4 1 .3 2 – 1 .4 7)

A g ed

8 0 + = 1 .1 7 (9 5 %

C I 1 .0 9– 1 .2 2 )

R R fo r ar m

fr ac tu re s (c o ld er

se as o n s v s m il d er

se as o n s) :

A g ed

6 5 – 7 9 = 1 .6 9 (9 5 %

C I 1 .5 6 – 1 .8 3 )

A g ed

8 0 + = 1 .3 0 (9 5 %

C I 1 .1 3– 1 .4 3 )

R R fo r h ip

fr ac tu re s (c o ld er

se as o n s v s m il d er

se as o n s) :

P eo p le ag ed

6 5 – 7 9 = 1 .2 7 (9 5 %

C I 1 .1 5 – 1 .3 7 )

P eo p le ag ed

8 0 + = 1 .0 8 (9 5 %

C I 1 .0 0– 1 .1 5 )

C am

p b el l et al .

(1 9 8 8 )

/C o h o rt st u d y /+

M o sg ie l, N ew

Z ea la n d /7 6 1

p er so n ag ed

70 +

p ar ti ci p at ed

in a p ro sp ec ti v e

st u dy

o f fa ll s

F al ls re co rd ed

by th e

p ar ti ci pa n ts fo r a p er io d o f

1 ye ar

N .A .

D ai ly

m in im

um te m pe ra tu re

S ig n if ic an t se as o n al v ar ia ti o n in

th e n u m b er

of fa ll s re p o rt ed

b y w o m en

li v in g in

th e

co m m u n it y (s p ri n g 2 3 .5 % ; su m m er

1 8 .9 % ;

au tu m n 22 .3 % ; w in te r 35 .2 % ) (P

< 0 .0 1 )

In si g n if ic an t se as o n al va ri at io n am

o n g w o m en

in re si d en ti al h o m es

2076 Int J Biometeorol (2018) 62:2073–2088

T ab

le 1

(c o n ti n u ed )

C it at io n /s tu d y

ty p e/ q u al it y

L o ca ti o n /p o p u la ti o n

O u tc o m e (s tu d y p er io d )

D ef in it io n

M et eo ro lo g ic al v ar ia b le s

R es u lt s

In d o or

(P < 0 .0 1 ) o r o u td o o r fa ll s (P

< 0 .0 5 )

si g n if ic an tl y in cr ea se

in w in te r fo r w o m en

b ut

n o t fo r m en

R R o f fa ll s:

T em

p er at u re ≤ 1 °C

v s > 1 °C

= 1 .5 3 (9 5 %

C I

1 .2 1 – 1 .8 4 )

C en te rs fo r D is ea se

C o n tr o l an d

P re v en ti o n (2 0 0 4 )/

co ho rt st u d y /+

U .S ./ P at ie n ts th at w er e tr ea te d

in U .S . ho sp it al em

er g en cy

d ep ar tm

en ts (E D s)

H o li d ay -d ec o ra ti n g -r el at ed

fa ll s

(1 N o v 2 0 0 0 to

3 1 Ja n 2 0 0 1 , 1

N o v 2 0 0 1 to

3 1 Ja n 2 0 0 2 , 1

N o v 2 0 0 2 to

3 1 Ja n 2 0 0 3 )

A p er so n d es ce n d ed

b ec au se

o f

th e fo rc e o f g ra v it y an d

st ru ck

a su rf ac e at th e sa m e

o r lo w er

le v el

N .A .

T h e o ve ra ll in ju ry

ra te w as

8 .1 p er

1 0 0 ,0 0 0

p o p u la ti o n (9 5 %

C I = 5 .9 – 1 0 .3 )

M o st in ju ri es

h ap p en ed

o n p er so n s ag ed

2 0 – 4 9 y ea rs (6 2 % )

M o re

m al es

(5 8 % ) su st ai ne d d ec o ra ti n g -r el at ed

fa ll s th an

fe m al es

(4 2 % )

4 3 %

o f fa ll s w er e fr o m

la d d er s an d 1 3 %

fr o m

ro o fs

M al es

w er e si g n if ic an tl y m o re li k el y th an

fe m al es

to fa ll fr o m la d d er s R R = 2 .4 (9 5 %

C I 1 .0 – 3 .7 ),

o r la d d er s an d ro o fs co m b in ed

R R = 3 .1

(9 5 %

C I 1 .8 – 4 .5 )

D ri ed g er

et al .

(2 0 1 6 )/ co h o rt

st ud y /+

C an ad a/ A ll p at ie n ts w h o w er e

se ve re ly

in ju re d an d

ad m it te d to

th e F o o th il ls

M ed ic al C en te r

F al l as

a d ir ec t re su lt o f th e

in st al la ti o n o f re si d en ti al

C h ri st m as

li gh ts

1 O ct to

2 4 D ec

o v er

a 1 0 -y ea r

p er io d (2 0 0 2 to

2 0 1 2 )

N .A .

A m b ie n t te m p er at u re

an d

ac ti v e sn o w fa ll

M o st pa ti en ts w er e m al e (9 5 % ) an d fa ll s fr o m

a la d de r w as

th e m ai n ca u se

(6 5 % )

N o st at is ti ca ll y si g n if ic an t co rr el at io n s fo u nd

b et w ee n se v er e in ju ri es

an d th e ti m e o f d ay ,

am b ie nt

te m p er at u re , ac ti v e sn o w fa ll o r th e

p ro x im

it y o f th e in ju ry

to C h ri st m as

D ay

(P > 0 .0 5 fo r ea ch )

G ev it z et al . (2 0 1 7 )

/C o h o rt st u d y /+

P h il ad el ph ia , U .S ./ F al l- re la te d

p at ie n ts v is it in g em

er g en cy

d ep ar tm

en t at 2 1

P h il ad el p h ia ar ea

h o sp it al s

E m er g en cy

d ep ar tm

en t v is it s

as so ci at ed

w it h fa ll s

(1 D ec

2 0 0 6 to

3 1 M ar

2 0 11 —

o n ly

D ec em

b er

to M ar ch )

K ey w o rd s Bf al l^ , Bf el l^ , Bs li p ^

o r Bt u m b le ^ id en ti fi ed

fr o m

th e em

er g en cy

d ep ar tm

en t

ch ie f co m p la in ts lo g s

A m b ie n t te m p er at u re (d ai ly

m in im

u m , m ax im

u m ,

an d av er ag e) , sn o w , ra in

an d fo g

H ig h -f al ld ay s h ap p en ed

m or e on

a w ee k d ay

th an

o n a w ee k en d (P

< 0 .0 0 1 )

2 0 %

o f fa ll -r el at ed

v is it s o cc u rr ed

b et w ee n

7 :0 0 am

an d 1 0 :5 9 am

an d 3 4 %

o cc u rr ed

b et w ee n 11 :0 0 am

an d 3 :4 9 p m

8 0 %

o f h ig h -f al l da y s w er e p re ce d ed

b y sn o w v s

3 7 %

o f co n tr o l d ay s (P

= 0 .0 0 2 ). H ig h er

av er ag e re co rd ed

pr ec ip it at io n in h ig h -f al ld ay s

(0 .7 , 9 5 %

C I 0 .3 – 1 .1 ) v s co n tr o l d ay s (0 .1 ,

9 5 %

C I 0 .1 – 0 .2 ).

O n ly

sn ow

w as

as so ci at ed

w it h fa ll -r el at ed

v is it s

in th e ad ju st ed

m o d el (A

d j. O R 5 = 1 3 .4 , 9 5 %

C I 2 .9 – 6 1 .5 )

G yl le n cr eu tz et al .

(2 0 1 5 )/ co h o rt

st ud y /+

S w ed en /P er so n s ag ed

6 5 +

at te n d ed

U m ea

U n iv er si ty

H o sp it al

P ed es tr ia n fa ll s in

a p u b li c ar ea

in th e ci ty

o f U m ea

m u n ic ip al it y

(J an

2 0 09

to A p r 2 0 11 )

N .A .

N .A .

T h e in ci de n ce

o f p ed es tr ia n fa ll s am

o n g el d er s

w as

2 7 /1 0 0 0 p er so n s p er

y ea r

M o st of

th em

(8 1 % ) w er e in ju re d d u ri n g w in te r

(N o v to

A p r)

Int J Biometeorol (2018) 62:2073–2088 2077

T ab

le 1

(c o n ti n u ed )

C it at io n /s tu d y

ty p e/ q u al it y

L o ca ti o n /p o p u la ti o n

O u tc o m e (s tu d y p er io d )

D ef in it io n

M et eo ro lo g ic al v ar ia b le s

R es u lt s

T h e m ai n ca u se

o f in ju ri es

w as

re la te d to

ic y

co n d it io n s (5 1 % )

H em

en w ay

an d

C o ld it z (1 9 9 0 )/

co ho rt st u d y / +

U .S ./ F em

al e re g is te re d n u rs e

ag ed

3 0 to

5 5 li v in g in

11 la rg e U .S . st at es

an d fa ll

d ea th s fo r 1 9 8 0

F al ld ea th st at is ti cs

fo r 1 9 8 0 ar e

ta k en

fr o m n at io n al m o rt al it y

ta p es , an d fr ac tu re

in fo rm

at io n co m es

fr o m

a lo n g it u d in al st u d y o f fe m al e

n u rs es

(1 9 8 2 – 1 98 4 )

N .A .

W ar m

st at es

w h er e th e

la rg es t ci ty

h ad

a m ea n

h ig h Ja n u ar y

te m p er at u re

o f 5 0 °F

o r

h ig h er , an d th e co ld er

st at es , w h er e th e m ea n

h ig h Ja n u ar y

te m p er at u re

d u ri n g th e

sa m e ti m e sp an

w as

b el o w 50

°F be tw ee n th e

y ea r o f 1 9 4 5 an d 1 9 8 0

T h e ag e- ad ju st ed

R R o f fa ll d ea th :

C o ld er

st at es

vs w ar m er

st at es = 1 .1 4 (9 5%

C I

1 .0 7 – 1 .2 0 )

F al ld ea th ra te s w er e h ig h er in w in te r m o n th s th an

in su m m er

m o n th s

F al l d ea th s in cr ea se

in w in te r m o n th s in

al l

re g io n s, b ut m o re ap p ar en ti n co ld er st at es

th an

in w ar m er

st at es

w h en

it co m pa re d to

su m m er

m o n th s (R R 1 .1 9 v s. 1 .1 2 )

T h e ag e- ad ju st ed

R R fo r fr ac tu re :

C o ld er

st at es

vs w ar m er

st at es = 1 .1 8 (9 5%

C I

1 .0 0 – 1 .3 9 )

R R o f fr ac tu re s (W

in te r v s su m m er ):

C o ld er

st at es = 1 .4 2 (P

= 0 .0 1 )

W ar m er

st at es = 1 .0 3 (P

= in si gn if ic an t)

Ja co bs en

et al .

(1 9 9 5 )/ co h o rt

st ud y /+

R o ch es te r, M in n es o ta ,

U .S ./ A ll R o ch es te r w o m en

ag ed

4 5 +

T h e d ai ly

oc cu rr en ce

of p ro x im

al fe m u r fr ac tu re

(1 Ja n 1 9 5 2 to

3 1 D ec

1 9 8 9 )

N .A .

S n o w , fr ee zi n g ra in ,

fr ee zi n g d ri zz le ,o r g la ze ,

h ig h w in d an d ra in

o r

d ri zz le

W o m en

ag ed

4 5 – 7 4 y ea rs , R R o f h ip

fr ac tu re :

D ay s w it h sn o w v s n o sn o w = 1 .4 1 (1 .1 0 – 1 .8 1 )

F re ez in g ra in

v s n o fr ee zi n g ra in = 1 .8 2

(1 .2 7 – 2 .6 2 )

W in te r v s su m m er = 1 .4 4 (9 5 %

C I 1 .0 6– 1 .9 6 )

W in te r v s su m m er (c o n tr o ll ed

fo r w ea th er ) = 1 .1 6

(9 5 %

C I 0 .8 1 – 1 .6 5 )

W o m en

ag ed

7 5 + , R R o f h ip

fr ac tu re :

ic e an d sn o w : in si g n if ic an t

W in te r v s su m m er

(c o n tr ol le d /u n co n tr o ll ed

fo r

w ea th er ) = 1 .1 6 (9 5 %

C I 0 .9 6 – 1 .4 0 )

L in

et al . (2 0 1 5 )/

co ho rt st u d y /+

T ai p ei C it y, T ai w an /P at ie n ts

re q u es te d em

er g en cy

m ed ic al se rv ic e (E M S )

T ra u m a pa ti en ts at te nd ed

to em

er g en cy

m ed ic al se rv ic e

(E M S )

(1 Ja n 2 0 0 9 to

3 1 D ec

2 0 1 0 )

F al ls re fe r to

an y in ju ry

re su lt in g fr o m

a p er so n

co m in g to

re st in ad v er te n tl y

o n th e g ro u n d o r fl oo r

H o u rl y m ea su re m en ts o f

te m p er at u re

(° C ),

p re ci p it at io n (m

m ),

su n sh in e (h ), h u m id it y

(% ), an d m ea n w in d

sp ee d (m

/s )

IR R 6 fo r li g h t, m o d er at e, an d h ea v y ra in vs

n o n e:

1 .0 7 , 1 .2 1 , an d 1 .3 2

1 h in cr ea se

in su n sh in e ex po su re

➔ 1 3 .2 %

in cr ea se

in th e ho u rl y in ci d en ce

(P < 0 .0 0 1 )

L at e fa ll an d w in te r v s sp ri n g an d su m m er :

in cr ea se d h o u rl y in ci d en ce s o f fa ll s (n o fi gu re s

p ro v id ed )

M o n d ay s (v s o th er

w ee k d ay s an d w ee k en d s)

IR R = 1 .0 9

N o as so ci at io n : w in d , h u m id it y an d te m p er at u re

L o p ez -S o to

et al .

(2 0 1 6 )/ co h o rt

st ud y /+

F er ra ra , It al y /P at ie n ts ag ed

6 5 + h o sp it al is ed

in th e fi v e

n o n te ac h in g p u b li c

h o sp it al s

In p at ie n ts fa ll

(I Ja n 2 0 1 0 to

31 D ec

2 0 1 3 )

F al l re fe rs to

th e co n se qu en ce

o f m o ve m en t u n in te n ti o n al ly

an d u n ex pe ct ed ly

to th e

g ro u n d fr o m

a h ig h er

p o si ti o n

N .A .

F al ls w er e si g n if ic an tl y m o re

fr eq u en t in

w in te r

an d sp ri n g (P

= 0 .0 0 3 ). F al ls w er e m o re

fr eq u en t o n F ri d ay s, S u n d ay s, an d M o n d ay s

(a b o u t 1 6 %

ea ch ) th an

T u es d ay s (a b o u t 12 % ),

b u t su ch

d if fe re n ce

w as

in si g n if ic an t.

2078 Int J Biometeorol (2018) 62:2073–2088

T ab

le 1

(c o n ti n u ed )

C it at io n /s tu d y

ty p e/ q u al it y

L o ca ti o n /p o p u la ti o n

O u tc o m e (s tu d y p er io d )

D ef in it io n

M et eo ro lo g ic al v ar ia b le s

R es u lt s

M aj o ri ty

o f fa ll s to o k p la ce

in th e p at ie n t’ s

h o sp it al ro o m

(7 2%

) an d b at hr o o m

(2 3 % ).

T h e m o st co m m o n ca u se

o f th e fa ll s w as

p at ie n t

in st ab il it y (3 2%

). M o st fa ll s o cc u rr ed

w h en

p at ie n ts w er e n o t w ea ri n g fo o tw ea r (4 5 % ) or

w ea ri n g in ap p ro p ri at e fo o tw ea r (3 9 % ).

H ig h er

fa ll o cc u rr ed

in th e n ig h t sh if t (9

p m

to 7

am ) (4 6%

) co m p ar ed

to ei th er

th e m o rn in g

sh if t (7

am to

2 p m ) (3 0 % ) o r af te rn o o n sh if t

(2 p m

to 9 p m ) (2 4 % ).

L u n d an d S h ea fo r

(1 9 8 5 )/ co h o rt

st ud y /+

U .S ./ E ld er s b ei ng

ad m it te d in

a sh o rt -s ta y co m m u n it y

h o sp it al

In p at ie n t fa ll s

(1 9 7 8 )

F al ls re p o rt ed

in in ci d en t

re p o rt s

N .A .

F al ls h ap p en ed

m o st fr eq u en tl y at b ed si d e (5 5 % )

P ea k fa ll ti m es

w er e o n th e ev en in g (3

p m

to 7

p m ) (2 4 % ) an d n ig h t sh if ts (3

am to

7 am

) (2 5 % ).

T u es d ay s an d T h u rs d ay s w er e th e d ay s o f th e

w ee k w it h 4 1 %

o f fa ll s h ap p en ed

T h e ac ti v it y m o st fr eq u en tl y as so ci at ed

w it h a fa ll

w as

g et ti n g in , o u t, an d ro ll in g o u t o f b ed

(4 7 % ).

R is k fa ct o rs o f in p at ie n t fa ll s in cl u d ed

se as o n

(f ro m S ep te m b er to N o v em

b er ), h av in g 3 + u n it

tr an sf er s, co g n it iv e im

p ai rm

en t, an d u se

of as si st iv e am

b u la to ry

d ev ic es , ta k in g v it am

in s,

ir o n , d iu re ti cs , h y p o te n si v es , an d /o r

an ti co n v u ls an ts

L u u k in en

et al .

(1 9 9 6 )/ co h o rt

st ud y /+

C it y o f O ul u ,

F in la n d /P o p u la ti o n ag ed

7 0 + an d li v in g at h o m e

A ll fa ll s, in cl u d in g th o se

th at

d id

n o t re su lt in

in ju ry , w er e

re co rd ed

(1 Ja n 1 9 9 1 to

3 1 D ec

1 9 9 3 )

N .A .

T h e d ai ly

m ea n

te m p er at u re s

IR R o f o u td o o r fa ll s fo r te m p er at u re

b el o w

− 2 0 °C

v s + 9 °C

= 4 .4 7

M ag o ta et al . (2 0 1 7 )/

co ho rt st u d y /+

F u k u o k a C it y, Ja p an /I n p at ie n t

fa ll er s in

a se co n d ar y

em er ge nc y m ed ic al fa ci li ty

In ci d en t re p o rt s re la te d to

in p at ie n t fa ll s fr o m

A pr il

2 0 1 0 to

M ar ch

2 0 1 4

A fa ll is an

ev en t in

w h ic h an

in d iv id u al fi n d s th em

se lv es

o n th e fl o o r u n in te n ti o n al ly

S u n ri se /s u n se t ti m es

an d

th e d ai ly

av er ag e

am b ie n t te m p er at u re s

F al l ra te o f n ig h t- ti m e w as

si g n if ic an tl y h ig h er

th an

th at o f d ay ti m e (1 .6

± 0 .8

v s.

1 .2

± 0 .7 /1 0 0 0 O B D s7 , P = 0 .0 01 )

A h ig h n u m b er o f fa ll s o cc u rr ed

d u ri ng

n ig h t- ti m e

an d d aw

n (2

a. m ., 5 a. m ., 6 a. m .a n d 7 a. m .) in

w in te r (e sp ec ia ll y, N o v em

be r, Ja n u ar y an d

F eb ru ar y ). N ig h t- ti m e le n g th (l o n g er in w in te r)

w as

si g n if ic an tl y re la te d to

an in cr ea se

in n ig h t- ti m e fa ll s (P

= 0 .0 4 7 ) S u n ri se -s un se t

ti m es

an d am

b ie n t te m p er at u re s w er e sh o w n

b y m o n th , b u t n o t u se d in th e an al y si s d ir ec tl y.

F al ls w er e m o st fr eq u en tl y re p o rt ed

at 7 a. m . in

F eb ru ar y (6 .7 /1 0 0 0 O B D s)

Int J Biometeorol (2018) 62:2073–2088 2079

T ab

le 1

(c o n ti n ue d )

C it at io n /s tu d y

ty p e/ q u al it y

L o ca ti o n /p o p u la ti o n

O u tc o m e (s tu d y p er io d )

D ef in it io n

M et eo ro lo g ic al v ar ia b le s

R es u lt s

N ig h t- ti m e le n g th

w as

si g n if ic an tl y re la te d to

an in cr ea se

in n ig h t- ti m e fa ll s (P

= 0 .0 4 7 )

T h e m o st fr eq u en t fa ll -r el at ed

b eh av io u r w as

to il et in g (5 6 .9 % )

M am

d an i an d U p sh ur

(2 0 0 1 )/ co h o rt

st u d y /+

O n ta ri o, C an ad a/ R es id en ts of

O n ta ri o

In ci de n ce

o f h o sp it al is at io n

re la te d to

fa ll (1

A p r 1 98 8 to

2 8 F eb

1 9 9 9 )

IC D -9 , C li n ic al M o di fi ca ti o n

co d es

(E 8 8 0 -E 8 8 8 )

N .A .

W ea k er

p at te rn s o f se as o n al v ar ia ti o n in

ad m is si o n s in

th o se

ag ed

60 +

P ea k ad m is si o n s in

D ec em

b er – A pr il (P

< 0 .0 1 )

an d an

u p w ar d tr en d o v er

ti m e (P

< 0 .0 1 )

T h e h o sp it al is at io n ra te am

o n g th o se

ag ed

6 0 +

w as

4 – 1 0 -f o ld

h ig h er

th an

o th er

ag e g ro u p s

M o d ar re s et al .

(2 0 1 2 )/ co h o rt

st u d y /+

M o nt re al , Q u eb ec

p ro vi n ce ,

C an ad a/ M o n tr ea l re g io n

re si d en ts ag ed

40 +

H o sp it al ad m is si o n w it h th e

m ai n ad m is si o n d ia g n o si s o f

a h ip

fr ac tu re

IC D -9

co de s

8 2 0 .X

(1 Ja n 1 9 9 3 to

3 1

D ec

2 0 0 4 )

R ec o rd s fr o m

in di v id u al s

w h o se

in ju ry

ex te rn al ca u se

w as

o th er

th an

Ba cc id en ta l

fa ll ^ (I C D -9

co d es

E 8 8 0 to

E 8 8 8 ) o r ac ci d en ts du e to

n at u ra l an d Be n v ir o n m en ta l

fa ct or s^

(I C D -9

co d es

E 9 0 0

to E 9 0 9 ) w er e ex cl u d ed

T em

p er at u re , p re ci p it at io n ,

sn o w ,w

in d an d su n sh in e

S ea so n al v ar ia ti o n n o te d fo r b o th

se x es .

F o r al l g en d er s an d ag e g ro u p s, h ip

fr ac tu re

ra te s

in cr ea se d w it h :

-d ec re as ed

te m p er at u re

(m ax im

u m , av er ag e,

m in im

u m ) (P

< 0 .0 1 )

-i n cr ea se d sn o w d ep th

an d th e n u m b er

o f sn ow

y d ay s (P

< 0 .0 1 )

-d ec re as ed

to ta l h o u rs o f su n sh in e (o r th e le n g th

o f a d ay ) (P

< 0 .0 5 fo r fe m al e ag ed

4 0– 7 4 ;

P < 0 .0 1 fo r o th er

re m ai n in g g ro u p s)

-d ec re as ed

ra in fa ll d ep th

(P < 0 .0 5 fo r fe m al es

ag ed

7 5 + ;P

< 0 .0 1 fo r o th er re m ai n in g g ro u p s)

-d ec re as ed

n u m b er

o f d ay s w it h ra in

(P < 0. 0 1 )

T h e m ax im

u m

an d m ea n w in d sp ee d sh o w ed

n o

re la ti on sh ip

w it h h ip

fr ac tu re

ra te s

M o n d o r et al . (2 0 1 5 )/

co h o rt st u d y / +

M o n t re al , C an ad a/ F al l- re la te d

in ju ri es

in d iv id u al w ho

p re se n te d to

an em

er ge n cy

d ep ar tm

en t

F al l re q u ir in g em

er ge n cy

d ep ar tm

en t ad m is si o n

(J an

1 1 9 9 8 to

D ec

3 1 2 0 0 6 )

U si n g a v al id at ed

se t o f

d ia g n o st ic an d p ro ce d u re

co d es

fo r fr ac tu re s,

su bl u x at io n s an d la ce ra ti o n s

p re se n te d (I C D -9 ) (c o d es

n o t

m en ti o n ed )

F re ez in g ra in

(i n cl u d in g

fr ee zi n g d ri zz le ) an d

al er ts fo r sn o w st o rm

s (i n cl u d in g h ea v y

sn o w fa ll s, sn o w sq u al ls ,

b lo w in g sn o w , b li zz ar d s

an d sn ow

st o rm

s, co m b in ed )

F re ez in g ra in

al er ts IR R (9 5 %

C I)

B o th

se x es

= 1 .2 0 (1 .0 8 – 1 .3 2 )

M al e = 1. 3 1 (1 .1 0 – 1 .5 6 )

S n o w st o rm

al er ts IR R (9 5 %

C I)

0 .8 9 (0 .8 0 – 0 .9 9 )

N o se as o n al d if fe re n ce

M o re n cy

et al .

(2 0 1 2 )/ co h o rt

st u d y /+

L av al an d M o n tr ea l Is la n d ,

C an ad a/ F al l p at ie n ts

re q u ir in g am

b u la n ce

se rv ic e

F al l re q u ir ed

am b u la n ce

in te rv en ti o n

(1 D ec

2 0 0 8 to

3 1 Ja n 2 0 0 9 )

N .A .

M ax im

um d ai ly

te m p er at u re , to ta l d ai ly

p re ci p it at io n (r ai n an d

sn o w ) an d p re se n ce

o f

fr ee zi ng

ra in

3 6 %

in d o o rs , 2 9 %

o u td o o rs , re m ai n in g

u n k n ow

n . M o re

y o u n g p eo p le th an

o ld

p eo p le

fe ll o ut d o o rs .

7 2 %

o f o u td o o r fa ll s w er e as so ci at ed

w it h ic e o r

sn ow

. E x ce ss

fa ll s w er e p re ce d ed

b y ra in

an d fo ll o w ed

b y fa ll in g te m p er at u re s, o r h ad

fr ee zi n g ra in .

O u td o o r fa ll s su b st an ti al ly

in cr ea se d at 1 to

3 d ay s af te r m et eo ro lo g ic al ev en ts fa v o ur ab le

to th e fo rm

at io n o f ic e o n si d ew

al k s.

2080 Int J Biometeorol (2018) 62:2073–2088

T ab

le 1

(c o n ti n u ed )

C it at io n /s tu d y

ty p e/ q u al it y

L o ca ti o n /p o p u la ti o n

O u tc o m e (s tu dy

p er io d )

D ef in it io n

M et eo ro lo g ic al v ar ia b le s

R es ul ts

P ar k er

an d M ar ti n

(1 99 4 )/ co h o rt

st u dy /+

U n it ed

K in g d o m /P at ie n ts

ad m it te d to

th e

B ir m in g h am

A cc id en t

H o sp it al

F al l w it h an

ac u te h ip

fr ac tu re

N .A .

G ro u n d fr o st an d m in im

u m

da il y te m p er at u re

S li g h t as so ci at io n b et w ee n d ay

o f fa ll an d th e

p re se n ce

o f g ro u n d fr o st (P

= 0 .0 4 )

N o se as o n al va ri an ce

(n o p o in t es ti m at e &

C I)

P ar k er

et al . (1 99 6 )/

co h o rt st u dy /+

U n it ed

K in g d o m /A ll p ri m ar y

ad m is si o n s to

P et er b o ro u g h

D is tr ic t H o sp it al

F al l w it h a h ip

fr ac tu re

(1 Ju n 1 9 92

to 3 1 M ay

1 9 9 5 )

N .A .

N .A .

M o st fa ll s o cc u rr ed

b et w ee n 09 0 0 an d n o o n

N o si g n if ic an tm

o n th ly v ar ia ti o n in th e n u m b er o f

fa ll s, in d o o r o r o u td o o r, an d no

o v er al ls ea so n al

v ar ia ti o n

P ip as

et al . (2 00 2 )/

co h o rt st u dy /+

N ew

Y o rk , U .S ./ A ll ad u lt

p at ie n ts p re se n te d to

th e

em er ge nc y d ep ar tm

en t at

th e re g io n al le v el 1 tr au m a

ce nt re

F al ls fr o m

ro o ft op

(J an

1 9 9 3

th ro u gh

D ec

1 9 9 6 )

In ju ry

co d es

(E -c o d es ) of

Bf al l

fr o m

b u il d in g o r o th er

st ru ct u re ^ (E 8 8 2 ), Bf al l fr om

a la d de r o r sc af fo ld ^ (E 8 8 1 ),

an d Bf al l fr o m

on e le v el to

an o th er ^ (E 8 84 .9 )

S n o w ac cu m u la ti o n

F al li n g fr o m a ro of to p af te r h ea v ie r sn o w fa ll s w as

si g n if ic an tl y as so ci at ed

w it h cl ea ri n g sn o w

(8 6 % ) (P

< 0 .0 0 1 )

S aa ri et al . (2 0 0 7 )/

co h o rt st u dy /+ +

Jy v as k y la , F in la n d /A ll th e

co m m u n it y li v in g re si d en ts

ag ed

7 5 an d 8 0 y ea rs

A cc id en ts an d in ju ri o u s fa ll s

du ri n g a 1 0 -y ea r fo ll o w -u p

pe ri o d

In fo rm

at io n w as

co d ed

ac co rd in g to

IC D -1 0

N .A .

N o se as o n al va ri at io n fo r b o th

in d o o r fa ll s

(P = 0 .6 7 3 ) an d o u td o o r fa ll s (P

= 0 .2 5 7)

A d j. R R 8 o f in ju ri o u s fa ll s:

O ld er

w o m en

v s m en

= 2 .1 1 (1 .5 2– 2 .9 1 )

O st eo ar th ri ti s = 1 .3 5 (1 .0 3 – 1 .7 6 )

N o ef fe ct fo r co ro n ar y h ea rt d is ea se s o r m o b il it y

li m it at io n

S m it h an d N el so n

(1 99 8 )/ co h o rt

st u dy /+

In d ia n a, U .S ./ P at ie n ts

ad m it te d to

th e M et h o d is t

H o sp it al o f In d ia n a b ec au se

o f a fa ll o n ic e

H ip

fr ac tu re

re q u ir in g E D

ad m is si on

(8 F eb

to 1 6 F eb ru ar y 1 9 9 4 )

F al l o n ic e as

in d ic at ed

as th e

ch ie f co m p la in t at

re g is tr at io n , in

th e n u rs in g

n o te , o r in

th e p h y si ci an

re co rd

N .A .

A n k le fr ac tu re s w er e m or e co m m o n th an

o th er

st u d ie s o f fa ll s af te r a sn o w st o rm

F ra ct u re

w as

si g n if ic an tl y re la te d to

in cr ea si n g

ag e (P

= 0 .0 0 0 1 )

F ra ct u re s an d in ju ri es

p ea k ed

o n th e 5 th

an d 6 th

d ay

af te r th e st o rm

S te v en s et al . (2 0 0 7 )/

co h o rt st u dy /+

U .S ./ O ld er

ad u lt s ag ed

6 5 +

y ea rs

F al l d ea th s fr o m

N at io n al

C en te r fo r H ea lt h S ta ti st ic s

an d n o n fa ta l fa ll -r el at ed

in -

ju ri es

fr o m

th e N at io n al

E le ct ro n ic In ju ry

S ur v ei ll an ce

S y st em

A ll

In ju ry

P ro gr am

(N E IS S -A

IP ) (D

ec 20 0 1 -N

o v 2 0 02 )

C au se

o f d ea th

d at a ar e co d ed

ac co rd in g to

IC D -1 0

U n in te n ti o n al fa ll -r el at ed

d ea th s w er e d ef in ed

as h av in g an

u n de rl y in g ca u se

o f d ea th

co d ed

W 0 0 -W

1 9

N o n fa ta l ca se s w er e

id en ti fi ed

in th e

N E IS S -A

IP d at a if

th ey

h ad

a p re ci p it at in g o r

im m ed ia te

ca u se

li st ed

as a Bf al l^

an d an

in te n t o f

in ju ry

d es ig n at ed

as Bu n in te n ti o n al /

u n d et er m in ed ^

C li m at e ca te g o ri es

w er e

de te rm

in ed

b y a st at e’ s

av er ag e te m p er at u re

o n

Ja n u ar y 2 0 0 1 . T h e

av er ag e te m p er at u re

o f

st at es

cl as si fi ed

as ha v in g a Bc ol d cl im

at e^

w as

≤ 3 2 °F

an d th e

st at es

cl as si fi ed

as ha v in g a Bw

ar m er

cl im

at e^

w it h

te m p er at u re s > 3 2 °F

N o se as o n al pa tt er n

F at al fa ll s: 9 .1 %

h ig h er

ra te in

co ld er

cl im

at es ,

re g ar d le ss

o f se as o n

Int J Biometeorol (2018) 62:2073–2088 2081

T ab

le 1

(c o n ti n u ed )

C it at io n /s tu d y

ty p e/ q u al it y

L o ca ti o n /p o p u la ti o n

O u tc o m e (s tu d y p er io d )

D ef in it io n

M et eo ro lo g ic al v ar ia b le s

R es u lt s

T u rn er

et al . (2 0 11 )/

co ho rt st u d y /+

S y d n ey

re g io n o f N ew

S o u th

W al es , A u st ra li a/ P at ie n ts

ag ed

6 5 +

D ai ly

fa ll -r el at ed

h ip

fr ac tu re

h o sp it al is at io n s (1

Ju l 1 9 9 8

to 3 1 D ec

2 0 0 4 )

IC D -1 0 -A

M co d e o f

S 7 2 .0 -S 7 2 .2 an d a fi rs t

ex te rn al ca u se

co d e of

W 0 0- W 1 9

M ea n d ai ly

te m p er at u re

R R (9 5 %

C I fo r a 1 °C

in cr ea se

in te m pe ra tu re ):

m en

7 5 – 8 4 : 0 .9 8 (0 .9 6 , 0 .9 9 )

m en

8 5 + : 0 .9 8 (0 .9 6 , 1 .0 0 )

w o m en

7 5 – 8 4 : 0 .9 9 (0 .9 8, 1 .0 0 )

w o m en

8 5 + : 0 .9 8 (0 .9 7 , 0 .9 9 )

R R fo r w ee k en d s v s w ee k d ay s

w o m en

6 5 – 7 4 : 0 .8 1

m en

8 5 + : 0 .8 9

V ik m an

et al . (2 0 11 )/

co ho rt st u d y /+

N o rt h er n S w ed en /P ar ti ci p an ts

h av in g re ce iv ed

re g u la r

m u n ic ip al it y h o m e h el p

se rv ic es

fo r lo ng

o r sh o rt

in te rv al s

F al ls d u ri n g a p ar ti cu la r y ea r

E v en ts in

w h ic h th e p er so n,

u n in te n ti o n al ly

an d

re g ar d le ss

o f ca u se , ca m e to

re st o n th e fl o o r o r an o th er

lo w er

le v el

M ea n d ay li g h t p h o to p er io d

(n u m b er

o f d ay li g h t

h o u rs fo r th e 1 5 th

d ay

o f

ea ch

m o n th ) o r w it h

m ea n te m p er at u re s

S ig n if ic an t co rr el at io n b et w ee n m o n th ly

in ci d en ce

an d d ay li g h t p h o to p er io d

(r 9 = − 0 .7 8 ; P = 0. 0 0 3 )

N o si g n if ic an t as so ci at io n w it h m ea n te m p er at u re

(r = − 0 .5 1 ; P = 0 .0 9 0 )

Y eu n g et al . (2 0 11 )/

co ho rt st u d y /+

H o n g K o n g /P o p u la ti o n ag ed

6 0 + at te n d in g A & E

D ep ar tm

en t in

a re g io n al

h o sp it al in

H o n g K o n g

F al l re qu ir in g A & E ad m is si on

(2 0 0 6 – 2 0 07 )

T h e A & E D ep ar tm

en t re co rd s

w h ic h w er e m ar k ed

w it h an

in d ex

Bf al l^

T h e m ea n d ai ly

m ax im

u m ,

m ea n an d m in im

u m

ai r

te m p er at u re

(° C ),

re la ti v e h u m id it y, an d

ra in fa ll in

ea ch

m o n th

M o re

fa ll s o cc u rr in g in

w in te r th an

sp ri n g o r

su m m er

(P < 0 .0 0 1 )

N u m b er

of fa ll s in cr ea se d w it h d ec re as ed

m ax im

u m

te m p er at u re

(r = − 0 .7 2) , m ea n

te m p er at u re

(r = − 0 .7 1 ), m in im

u m

te m p er at u re

(r = − 0 .7 0 ) &

re la ti v e h u m id it y

(r = − 0 .7 4 ) b u t n o t ra in fa ll

L ar g er p ro p o rt io n o f fa ll s o cc u rr ed

am o n g p eo p le

li v in g in o ld ag e h o m e (2 0 .5 % ) th an

o w n h o m e

(1 1 .8 % ) du ri n g p ea k se as o n (w

in te r an d

au tu m n )

L o w er

li m b w ea k ne ss

w as

si g n if ic an tl y

as so ci at ed

w it h fa ll o cc u rr en ce

in p ea k se as on

am o n g al l p re d is p o si n g fa ct o rs (P

< 0 .0 5 )

S im

il ar

p ro p o rt io n s o f fa ll s in do o rs (5 3 % ) an d

o u td o o rs (4 7 % )

1 N o t av ai la b le ; 2 P v al u e;

3 R at e ra ti o ; 4 C o n fi d en ce

in te rv al s;

5 A d ju st ed

o dd s ra ti o ; 6 In ci d en ce

ra te ra ti o ; 7 O cc u p ie d be d -d ay ; 8 A d ju st ed

ra te ra ti o;

9 P ea rs o n co rr el at io n co ef fi ci en ts

2082 Int J Biometeorol (2018) 62:2073–2088

Study periods and unit of analysis

The study periods vary across all the studies, with the earliest data coming from 1952 in United States (Jacobsen et al. 1995). It was also the study with longest study durations which records the daily incidence over 37 years. The latest study used the data from 2010 to 2014 in Japan (Magota et al. 2017). Study period durations ranged from a week (Smith and Nelson 1998) to 10 years or more (Bergstrom et al. 2008; Driedger et al. 2016; Jacobsen et al. 1995; Mamdani and Upshur 2001; Modarres et al. 2012; Saari et al. 2007). Many studies used daily data (Campbell et al. 1988; Driedger et al. 2016; Gevitz et al. 2017; Jacobsen et al. 1995; Lin et al. 2015; Luukinen et al. 1996; Mondor et al. 2015; Morency et al. 2012; Parker and Martin 1994; Pipas et al. 2002; Smith and Nelson 1998; Turner et al. 2011) or monthly data for analysis (Aharonoff et al. 1998; Arbes and Berzlanovich 2015; Bergstrom et al. 2008; Bulajic-Kopjar 2000; Gyllencreutz et al. 2015; Magota et al. 2017; Mamdani and Upshur 2001; Modarres et al. 2012; Parker et al. 1996; Vikman et al. 2011; Yeung et al. 2011) whereas seasonal data (Centers for Disease Control and Prevention 2004; Hemenway and Colditz 1990; Lund and Sheafor 1985; Saari et al. 2007; Stevens et al. 2007) were less commonly used for analysis. There was two studies analysing hourly patterns of falls (Lopez-Soto et al. 2016; Lund and Sheafor 1985; Magota et al. 2017).

Subject characteristics

Nine studies focused on elders aged 65 or more (Aharonoff et al. 1998; Bulajic-Kopjar 2000; Gyllencreutz et al. 2015; Lopez-Soto et al. 2016; Lund and Sheafor 1985; Mondor et al. 2015; Stevens et al. 2007; Turner et al. 2011; Vikman et al. 2011). Another nine had various age cut-off points rang- ing from people with an age of 16 or above (Pipas et al. 2002), 40 or above (Modarres et al. 2012), 45 or above (Jacobsen et al. 1995), 50 or above (Bergstrom et al. 2008), 60 or above (Parker and Martin 1994; Yeung et al. 2011), 70 or more (Campbell et al. 1988; Luukinen et al. 1996) or even to 75 or more (Saari et al. 2007). Eleven of them specified no age boundaries (Arbes and Berzlanovich 2015; Centers for Disease Control and Prevention 2004; Driedger et al. 2016; Gevitz et al. 2017; Hemenway and Colditz 1990; Lin et al. 2015; Magota et al. 2017; Mamdani and Upshur 2001; Morency et al. 2012; Parker et al. 1996; Smith and Nelson 1998).

Identification of fall cases and outcome of fall

Some of the fall cases were identified with the aid of International Classification of Diseases (ICD) external cause code indicating an unintentional fall during admission (Bulajic-Kopjar 2000; Mamdani and Upshur 2001;

Modarres et al. 2012; Mondor et al. 2015; Pipas et al. 2002; Saari et al. 2007; Stevens et al. 2007; Turner et al. 2011). Eight studies had their own definition for fall (Arbes and Berzlanovich 2015; Centers for Disease Control and Prevention 2004; Gevitz et al. 2017; Lin et al. 2015; Lopez- Soto et al. 2016; Magota et al. 2017; Morency et al. 2012; Vikman et al. 2011). Most studies specified the exclusion criteria such as fractures more distal on the proximal femur (Jacobsen et al. 1995), motor vehicle crashes (Bulajic-Kopjar 2000; Gyllencreutz et al. 2015; Jacobsen et al. 1995; Morency et al. 2012), intentional falls (Morency et al. 2012), fall from height (Jacobsen et al. 1995), pathologic cause (Aharonoff et al. 1998; Jacobsen et al. 1995), fall in hospital (Parker and Martin 1994), fall in non-snow months (Pipas et al. 2002), unknown injury mechanism (Driedger et al. 2016; Pipas et al. 2002), no history of a fall (Parker and Martin 1994), no fall date and time, inpatient falls that occurred in the emer- gency and surgical departments (Lopez-Soto et al. 2016) and occupational injuries (Bulajic-Kopjar 2000; Driedger et al. 2016). There were studies included high energy falls (fall > 5 m) (Arbes and Berzlanovich 2015; Bergstrom et al. 2008) and vehicle-related incidents (Bergstrom et al. 2008). All of the selected studies focused on the incidences of fall requiring medical attention, except two emphasised on the fall mortality as well (Hemenway and Colditz 1990; Stevens et al. 2007). There was one study investigated on fatal falls only (Arbes and Berzlanovich 2015).

Demographic factors analysed

Demographic data were classified into two factors, namely gender and age. There were eight studies mentioning the age effect (Aharonoff et al. 1998; Hemenway and Colditz 1990; Mamdani and Upshur 2001; Mondor et al. 2015; Smith and Nelson 1998; Stevens et al. 2007; Turner et al. 2011; Vikman et al. 2011). The percentage of falls tended to be higher in people aged over 75 years (Mondor et al. 2015). People aged over 85 years had the highest fall-related hip fracture hospitalisation rates (Turner et al. 2011) while one study found that nonfatal fall-related injury rates did not rise sharply for those 85 years and older as compared to other age groups (Stevens et al. 2007). For fatal fall, the number rose sharply with increasing age (Hemenway and Colditz 1990; Stevens et al. 2007). It was found that women over 85 had over 30 times the likelihood of fall death as women in their late 60s (Hemenway and Colditz 1990).

There were six studies examining gender and most results indicated that women were more likely to fall than men (Bergstrom et al. 2008; Mondor et al. 2015; Saari et al. 2007; Smith and Nelson 1998; Stevens et al. 2007; Vikman et al. 2011). On average, there were more fall-related hip frac- ture per day for females than males (Turner et al. 2011). Women had approximately double the risk for an injurious

Int J Biometeorol (2018) 62:2073–2088 2083

fall than men (Saari et al. 2007). Bergstrom et al. (2008) men- tioned that the absolute number of fractures in male were decreasing after puberty whereas the number of fractures in female increased after age 40 and continued until age 80. The reason for the difference was suggested to be due to hormonal changes and related loss of muscle strength close to meno- pausal age for women (Saari et al. 2007). Yet, the results were reverse for inpatient falls and fatal falls. Hospital falls were significantly more common in men than women (Lopez-Soto et al. 2016). In addition, fatal fall rates were significantly higher among men than women overall (Stevens et al. 2007; Arbes and Berzlanovich 2015). There was only one excep- tional study found that the percentages of men and women who fell were virtually identical (Vikman et al. 2011).

Meteorological factors analysed

A wide variety range of meteorological variables has been studied. Temperature (13 studies) was the most commonly investigated meteorological factors. Other investigated factors included ice and snow (9 studies), precipitation (6 studies), daylight hours (3 studies), wind speed (3 studies), freezing rain (3 studies), humidity (2 studies), ground frost (1 study) and fog (1 study). Besides the meteorological factors, there were three environmental factors identified in the studies namely, seasonal factors (20 studies), locations of fall (in- door/outdoor) (6 studies) and time of the day (8 studies). Significant associations between the incidence or mortality of falls were found among a few climatic risk factors namely seasonal effect, temperature and ice/snow and were summarised as below.

Seasonal effect

Season was one of the important predictors of fall risk which was studied by 20 selected studies. The findings were con- flicting though. Seven of them found the rate of fall-related injuries did not differ seasonally (Aharonoff et al. 1998; Bergstrom et al. 2008; Mondor et al. 2015; Parker and Martin 1994; Parker et al. 1996; Saari et al. 2007; Stevens et al. 2007). Even fatal fall rates did not show any seasonal patterns neither (Stevens et al. 2007). A study showed patterns of seasonal variation in admissions varied among different age groups. The 0–9 year age group experienced increased fall- related hospital admissions during typically warmer months (May through November) whereas the 30–59-year age group experienced increases in the typically colder months (December to April). Weaker patterns were noted in the 60 and older age group with peaks occurring in the typically colder months (December to April) (Mamdani and Upshur 2001).

Another 12 studies identified apparent seasonal variations (Arbes and Berzlanovich 2015; Bulajic-Kopjar 2000;

Campbell et al. 1988; Gyllencreutz et al. 2015; Hemenway and Colditz 1990; Jacobsen et al. 1995; Lin et al. 2015; Lopez-Soto et al. 2016; Lund and Sheafor 1985; Magota et al. 2017; Modarres et al. 2012; Yeung et al. 2011). Most of them found that the incidence of falls or hip fracture rates increased only in late fall and winter. This was also the case in subtropical places like Taiwan (Lin et al. 2015) and Hong Kong (Yeung et al. 2011). A study revealed that a seasonal pattern was apparent for both younger and older women, but was considerably more pronounced among the younger wom- en. The rationale suggested behind is that older women would be expected to spend less time outdoors, reducing their expo- sure to inclement weather (Jacobsen et al. 1995). Similarly, one study indicated that more fall-related visits were found among adults aged 18–64 than elders aged 65 or above in winter. They suggested that elders were more well protected from inclement winter weather in general (Gevitz et al. 2017).

In Japan & the United States, inpatient falls showed an increase in fall or winter months (Lund and Sheafor 1985; Magota et al. 2017) while in Italy, hospital falls also happened more frequently in spring besides winter months (Lopez-Soto et al. 2016). The rise in inpatient falls in winter might be due to delirium, which is found commonly in older hospitalised pa- tients in winter (Gallerani and Manfredini 2013). Fall deaths, however, were inconclusive. One study noted an increase in winter (Hemenway and Colditz 1990) whereas another study found fatal fall incidence was higher in summer months than in winter months (Arbes and Berzlanovich 2015).

Temperature

Ambient temperature is one of the most common meteorolog- ical risk factors analysed in our reviewed studies (13 studies). However, these studies defined temperature differently. Studies analysed the data on a monthly, daily or even an hour basis. Two studies analysed the daily mean temperature (Luukinen et al. 1996; Turner et al. 2011), two analysed the daily minimum temperature (Campbell et al. 1988; Parker and Martin 1994) and one used the daily maximum temperature (Morency et al. 2012). Another study utilised all daily mini- mums, maximums and averages (Gevitz et al. 2017). For monthly data, there was one using the monthly mean temper- ature (Vikman et al. 2011) while another two studies investi- gated the effect of maximum, minimum and mean temperature within a monthly time scale (Modarres et al. 2012; Yeung et al. 2011). There were two studies compared warm states and colder states in their countries based on their defined temperature range (Hemenway and Colditz 1990; Stevens et al. 2007). One study used hourly temperature as measure- ment unit (Lin et al. 2015). Eight studies reported a significant negative association between fall outcomes and temperature (Campbell et al. 1988; Hemenway and Colditz 1990; Luukinen et al. 1996; Modarres et al. 2012; Morency et al.

2084 Int J Biometeorol (2018) 62:2073–2088

2012; Stevens et al. 2007; Turner et al. 2011; Yeung et al. 2011) i.e. lower air temperature was associated with more fall-related incidences or deaths. A study emphasised that rain followed by a drop in temperature might be an important meteorological factor for outdoor falls (Morency et al. 2012). Five studies did not report any significant association (Driedger et al. 2016; Gevitz et al. 2017; Lin et al. 2015; Parker and Martin 1994; Vikman et al. 2011). The small sam- ple size and short observation time were common among studies with insignificant results.

Snow cover

Ice/snow factor was well studied in many northern countries where all of them shared a freezing cold weather condition in winter. Seven studies found that fall injuries on ice/snow were common on the days with snow or a few days after snow (Gevitz et al. 2017; Gyllencreutz et al. 2015; Jacobsen et al. 1995; Mondor et al. 2015; Morency et al. 2012; Pipas et al. 2002; Smith and Nelson 1998). Peak injury events were likely to present several days after an ice/snow storm following the melting of them as temperature rose and time passed elevates the slipping chances (Smith and Nelson 1998). In Lewis and Lasater’s study, ice-related falls peaked between the fifth and eighth days of icy conditions (Lewis and Lasater 1994). Climbing onto a rooftop to clear significant accumulations of snow (> 12 in.) was reported as an activity linked with an elevated risk of fall-related injury (Pipas et al. 2002). Since Christmas occurred in winter, one study investigated the link- age with decoration activity. It reported no statistically signif- icant correlations between active snowfall and fall as a direct result of installation of residential Christmas lights (Driedger et al. 2016).

There were different findings with age group in these stud- ies. The snow depth and the number of snowy days show a positive significant correlation with hip fracture rates for all age groups and genders (Mondor et al. 2015). On the other hand, one study revealed that positive correlation between hip fracture and increased number of snowy days was observed among the women aged 45–74 years but not the women aged 75 years and older. The reason suggested behind was that old women would be expected to spend less time outdoors, reduc- ing their exposure to inclement weather (Jacobsen et al. 1995).

Other environmental risk factors

Location of fall

Where the fall took place was another factor investigated in six studies with diverse findings. Six studies reported that the majority of falls and hip fractures occur at home (Aharonoff et al. 1998; Bergstrom et al. 2008; Campbell et al. 1988; Parker et al. 1996; Saari et al. 2007; Yeung et al. 2011).

However, two studies showed that the distribution of fractures occurring indoor and outdoor was fairly equal (Bergstrom et al. 2008; Yeung et al. 2011). Of all fractures, 46% occurred indoors and 44% outdoors (Bergstrom et al. 2008).

Aharonoff et al. (1998) further explained that fractures par- ticularly occur in people who were older, less healthy, and had poorer ambulatory functions. This may suggest that why many previous studies showed that older people are likely to fall at home. Extrinsic mechanisms are more common in younger active people, and therefore the falls happen more often outdoors whereas falls due to intrinsic or unknown mechanisms occur most likely in the less active older age group (Ryynanen et al. 1991).

Time of the day

There were eight studies evaluating the time of the day in which the fall occurred (Aharonoff et al. 1998; Gevitz et al. 2017; Lin et al. 2015; Lopez-Soto et al. 2016; Lund and Sheafor 1985; Magota et al. 2017; Parker et al. 1996; Turner et al. 2011). A study reported significantly more elders fell during daylight hours with a peak in the afternoon. They sup- posed that they are likely to fall due to the increased activity of meal preparation (Aharonoff et al. 1998). Parker et al. (1996) also found higher incidence of falls happened in the morning and reported a lowest incidence during overnight. He attribut- ed this finding to the effects of long-acting benzodiazepine drugs and hypotensive agents and the activities of getting up, dressing and starting the day. For adults’ population, many falls tended to occur in the morning during work- or business- related commute (Gevitz et al. 2017). On the other hand, in- patient falls were more frequently observed during night-time and dawn (Lopez-Soto et al. 2016; Lund and Sheafor 1985; Magota et al. 2017). Some explained the association by the dark and cold conditions in the mornings (Magota et al. 2017), while explained by the change in the patient activities and the enhanced demands of nursing staff (Lopez-Soto et al. 2016; Lund and Sheafor 1985). Ryynanen et al. summarised that intrinsic or unknown mechanisms of falling commonly took place in the evening or at night as commonly caused by certain kinds of health diseases or disorders such as dementia and orthostatic hypotension while extrinsic mechanisms were common for the fall took place in the morning or in the after- noon since this type of fall depends on the time of day and amount of activity. This rule is applicable to all population (1991). This sounds reasonable as the younger active group are less vulnerable to health-related falls.

Interestingly, fall occurrences tended to increase on Mondays (Lin et al. 2015). Atherton and colleagues suggested that elderly people engaged in more outdoor activity during the initial days of the week (Atherton et al. 2005). Also, Turner et al. found lower fall hospitalisations on weekends compared to weekdays. They suggested whether this is again

Int J Biometeorol (2018) 62:2073–2088 2085

related to the activity levels or more social and family support (2011). For inpatient falls, no consistent pattern was being observed. More falls were noticed on Fridays, Sundays and Mondays in one study (Lopez-Soto et al. 2016) while another study found Tuesdays and Thursday to be the most frequent fall days of the week (Lund and Sheafor 1985).

Discussion

This systematic review summarised the past relevant studies on the association between fall-related injuries and the mete- orological factors. Even though those studies were conducted worldwide, there are some consistent findings among a few climatic factors investigated, in particular, lower ambient tem- perature and snow/ice cover. Other factors seem inconclusive and require more supporting evidence before reaching a consensus.

It is expected that lower ambient temperature will result in more fall occurrences. Although there is no agreed seasonal effect found in winter, the inclement cold weather will in- crease the fall risks. There are a few possible mechanisms underlying the cold weather that many studies suggesting most elderly’s falls related to. Impairment of judgement and coordination that were associated with slight falls in body temperature may contribute to the likeliness to fall among undernutrition old ages people (Bastow et al. 1983). Lower ambient temperatures were found to affect blood pressure and haemodynamics (Collins et al. 1985; Keatinge et al. 1984) and diminish dexterity (Riley and Cochran 1984). Hypothermia associated with low body temperatures and slow reaction time increase the risk of falls in cold weather (Atherton et al. 2005). This may be the reason why so many elders fell indoors even though they did not expose to the freezing cold weather out- side. Another possibility is that elderly person tends to put on extra clothes under cold weather, which in turn may make them more clumsy and prone to fall injury (Douglas et al. 2000). There is another explanation that colder temperatures may also decrease physical activity which result in impaired conditioning and ultimately subsequent bone fragility (Bergstralh et al. 1990).

For ice and snow, we postulate that the snow/ice cover increases the risk of slipping during the winter. It is foresee- able that people would sustain to fall more likely on ice/snow due to the extremely slippery conditions which led to a low coefficient of friction and increased torque as the victims slipped while weighting the extremity (Smith and Nelson 1998). Thus, ice/snow is a significant predictor of fall. In advance of a fall, proactive alerts to forecasts of snow and ice storms should be initiated. Weather condition application in smart phones nowadays might be an innovative idea to warn citizens the changes in the environment or road condi- tions rapidly (Gyllencreutz et al. 2015). Reporting emphasised

on the health risks associated with rooftop snow clearing might be helpful to reduce the incidence of falls too (Pipas et al. 2002). Adequate snow removal and de-icing is another effective method to reduce pedestrian falls (Gevitz et al. 2017; Morency et al. 2012). However, this factor is not applicable to relatively warm places where they never snowed, nor to places when residents are not living in individual apartments.

In addition to the weather variables, there were inconsistent findings on other environmental risk factors such as the loca- tions of fall. There were some questions raised on whether the places of fall determine the mechanism of fall, i.e. whether the fallers fail to keep their home warm or they went out to the cold environment. It was suggested to include fall locations as one of the determinants in fall injury.

It is likely that the environmental risk factors of fall can be preventable. When it comes to the days that the prevail- ing weather is going to be freezing cold or snowing, the government and related agencies should adopt timely pre- ventive measures regarding the adverse weather conditions like issue warning signals, enhance health education, pro- mote speedy removal of ice/snow, provide adequately heat- ed houses and distribute adequate amount of necessity to maintain an adequate temperature e.g. clothing and blan- kets. Innovative clothing is another good alternative for elders to fight against cold (Yeung et al. 2011). More at- tention should be paid to elders for the risk factors present at the home environment. Additionally, multidisciplinary team can join hands to improve the fall situation like oc- cupational therapist can help improve the environment of indoor living. Physiotherapist can conduct frequent physi- cal fitness training in the community or at home to enhance the muscle strength and endurance of bones and muscles of the public to combat the risk of falling due to inactivity. By taking these proactive preventive measures, it is hoped that the fall risks can largely be reduced.

There has been limited research on this research area especially in the Asian countries. The above findings might not be applicable to a subtropical or tropical region due to the difference in living environment and population characteristics. It is therefore requiring another large-scale research study with more weather and environmental fac- tors involved and with longer time period to reveal the linkage of fall and environment factors (Aharonoff et al. 1998; American Geriatrics Society et al. 2001).

Conclusion

In summary, there existed a few meteorological factors includ- ing the lower air temperature and the presence of ice/snow cover that were likely to have an impact on fall. It is of para- mount importance to make all the concerned parties to be aware of the fall circumstances.

2086 Int J Biometeorol (2018) 62:2073–2088

Acknowledgements This study is not funded by any projects.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict of interest.

References

Aharonoff GB, Dennis MG, Elshinawy A, Zuckerman JD, Koval KJ (1998) Circumstances of falls causing hip fractures in the elderly. Clin Orthop Relat Res 348:10–14

American Geriatrics Society, British Geriatrics Society, & American Academy of Orthopaedic Surgeons Panel on Falls Prevention (2001) Guideline for the prevention of falls in older persons. J Am Geriatr Soc 49(5):664–672

Arbes S, Berzlanovich A (2015) Injury pattern in correlation with the height of fatal falls. Wien Klin Wochenschr 127(1–2):57–61

Atherton WG, Harper WM, Abrams KR (2005) A year’s trauma admis- sions and the effect of the weather. Injury 36(1):40–46

Bastow MD, Rawlings J, Allison SP (1983) Undernutrition, hypothermia, and injury in elderly women with fractured femur: an injury re- sponse to altered metabolism? Lancet 1(8317):143–146

Berg WP, Alessio HM, Mills EM, Tong C (1997) Circumstances and consequences of falls in independent community-dwelling older adults. Age Ageing 26(4):261–268

Bergstralh EJ, Sinaki M, Offord KP, Wahner HW, Melton LJ (1990) Effect of season on physical activity score, back extensor muscle strength, and lumbar bone mineral density. J Bone Miner Res 5(4): 371–377

Bergstrom U, Bjornstig U, Stenlund H, Jonsson H, Svensson O (2008) Fracture mechanisms and fracture pattern in men and women aged 50 years and older: a study of a 12-year population-based injury register, Umea, Sweden. Osteoporos Int 19(9):1267–1273

Bulajic-Kopjar M (2000) Seasonal variations in incidence of fractures among elderly people. Inj Prev 6(1):16–19

Campbell AJ, Spears GF, Borrie MJ, Fitzgerald JL (1988) Falls, elderly women and the cold. Gerontology 34(4):205–208

Carpenter CR (2009) Will my patient fall? Ann Emerg Med 53(3):398– 400

Centers for Disease Control & Prevention (2004) Fall-related injuries during the holiday season—United States, 2000-2003. Morb Mortal Wkly Rep 53(48):1127–1129

Chu LW, Chi I, Chiu AY (2005) Incidence and predictors of falls in the Chinese elderly. Ann Acad Med Singap 34(1):60–72

Chu LW, Chi I, Chiu AY (2007) Falls and fall-related injuries in community-dwelling elderly persons in Hong Kong: a study on risk factors, functional decline, and health services utilization after falls. Hong Kong Med J 13(Suppl 1):S8–S12

Collins KJ, Easton JC, Belfield-Smith H, Exton-Smith AN, Pluck RA (1985) Effects of age on body temperature and blood pressure in cold environments. Clin Sci (Lond) 69(4):465–470

Douglas S, Bunyan A, Chiu KH, Twaddle B, Maffulli N (2000) Seasonal variation of hip fracture at three latitudes. Injury 31(1):11–19

Driedger MR, Gupta A, Wells B, Dixon E, Ball CG (2016) BOh the weather outside is frightful^: severe injury secondary to falls while installing residential Christmas lights. Injury 47(1):277–279

Gallerani M, Manfredini R (2013) Seasonal variation in the occurrence of delirium in patients admitted to medical units of a general hospital in Italy. Acta Neuropsychiatr 25(3):179–183

Gevitz K, Madera R, Newbern C, Lojo J, Johnson CC (2017) Risk of fall- related injury due to adverse weather events, Philadelphia, Pennsylvania, 2006-2011. Public Health Rep 132(1_suppl):53s–58s

Gyllencreutz L, Bjornstig J, Rolfsman E, Saveman BI (2015) Outdoor pedestrian fall-related injuries among Swedish senior citizens—in- juries and preventive strategies. Scand J Caring Sci 29(2):225–233

Hemenway D, Colditz GA (1990) The effect of climate on fractures and deaths due to falls among white women. Accid Anal Prev 22(1):59– 65

Ip D, Ip FK (2006) Elderly patients with two episodes of fragility hip fractures form a special subgroup. Journal of Orthop Surg (Hong Kong) 14(3):245–248

Jacobsen SJ, Sargent DJ, Atkinson EJ, O'Fallon WM, Melton LJ (1995) Population-based study of the contribution of weather to hip fracture seasonality. Am J Epidemiol 141(1):79–83

Kannus P, Parkkari J, Koskinen S, Niemi S, Palvanen M, Jarvinen M, Vuori I (1999) Fall-induced injuries and deaths among older adults. J Am Med Assoc 281(20):1895–1899

Keatinge WR, Coleshaw SR, Cotter F, Mattock M, Murphy M, Chelliah R (1984) Increases in platelet and red cell counts, blood viscosity, and arterial pressure during mild surface cooling: factors in mortality from coronary and cerebral thrombosis in winter. Br Med J 289(6456):1405–1408

Kojima S, Furuna T, Ikeda N, Nakamura M, Sawada Y (2008) Falls among community-dwelling elderly people of Hokkaido, Japan. Geriatr Gerontol Int 8(4):272–277

Lewis LM, Lasater LC (1994) Frequency, distribution, and management of injuries due to an ice storm in a large metropolitan area. South Med J 87(2):174–178

Lilley JM, Arie T, Chilvers CE (1995) Accidents involving older people: a review of the literature. Age Ageing 24(4):346–365

Lin LW, Lin HY, Hsu CY, Rau HH, Chen PL (2015) Effect of weather and time on trauma events determined using emergency medical service registry data. Injury 46(9):1814–1820

Liu Y, Liu WW, Wang LL (2015) Lao Nian Ren Die Dao De Wei Xian Yin Su Ji Hu Li Gan Yu [Risk factors of fall among elderly]. J Mod Med Health 31(5):688–690

Lopez-Soto PJ, Smolensky MH, Sackett-Lundeen LL, De Giorgi A, Rodriguez-Borrego MA, Manfredini R, . . . Fabbian F (2016) Temporal patterns of in-hospital falls of elderly patients. Nurs Res 65(6)435–445

Lord SR, Sherrington C, Menz HB (2001) Falls in older people : risk factors and strategies for prevention. Cambridge University Press, Cambridge

Lord S, Sherrington C, Menz H, Close J (2007) Falls in older people. Risk factors and strategies for prevention, 2nd edn. Cambridge University Press, Cambridge

Lund C, Sheafor ML (1985) Is your patient about to fall? J Gerontol Nurs 11(4):37–41

Luukinen H, Koski K, Kivela SL (1996) The relationship between out- door temperature and the frequency of falls among the elderly in Finland. J Epidemiol Community Health 50(1):107

Magota C, Sawatari H, Ando S-i, Nishizaka MK, Tanaka K, Horikoshi K, . . . Chishaki A (2017) Seasonal ambient changes influence inpatient falls. Age Ageing, 46(3)513–517

Mamdani MM, Upshur RE (2001) Fall-related hospitalizations: what's in season? Can J Public Health 92(2):113–116

Mertz KJ, Lee DC, Sui X, Powell KE, Blair SN (2010) Falls among adults: the association of cardiorespiratory fitness and physical ac- tivity with walking-related falls. Am J Prev Med 39(1):15–24

Modarres R, Ouarda TB, Vanasse A, Orzanco MG, Gosselin P (2012) Modeling seasonal variation of hip fracture in Montreal, Canada. Bone 50(4):909–916

Moher D, Liberati A, Tetzlaff J, Altman D, Group, P (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 151(4):264–269

Mondor L, Charland K, Verma A, Buckeridge DL (2015) Weather warn- ings predict fall-related injuries among older adults. Age Ageing 44(3):403–408

Int J Biometeorol (2018) 62:2073–2088 2087

Morency P, Voyer C, Burrows S, Goudreau S (2012) Outdoor falls in an urban context: winter weather impacts and geographical variations. Can J Public Health 103(3):218–222

Nagurney JT, Borczuk P, Thomas SH (1998) Elderly patients with closed head trauma after a fall: mechanisms and outcomes. J Emerg Med 16(5):709–713

National Center for Injury Prevention and Control (2004). Web-based injury statistics query and reporting system. Retrieved from http:// www.cdc.gov/ncipc/wisqars

Parker MJ, Martin S (1994) Falls, hip fractures and the weather. Eur J Epidemiol 10(4):441–442

Parker MJ, Twemlow TR, Pryor GA (1996) Environmental hazards and hip fractures. Age Ageing 25(4):322–325

Pipas L, Schaefer N, Brown LH (2002) Falls from rooftops after heavy snowfalls: the risks of snow clearing activities. Am J Emerg Med 20(7):635–637

Riley MW, Cochran DJ (1984) Dexterity performance and reduced am- bient temperature. Hum Factors 26(2):207–214

Roman Ortiz C, Tenias JM, Estarlich M, Ballester F (2015) Systematic review of the association between climate and hip fractures. Int J Biometeorol 59(10):1511–1522

Rubenstein LZ, Josephson KR (2002) The epidemiology of falls and syncope. Clinics in Geriatrics Medicine. W.B Saunders Co., Philadelphia

Ryynanen OP, Kivela SL, Honkanen R (1991) Times, places, and mech- anisms of falls among the elderly. Z Gerontol 24(3):154–161

Saari P, Heikkinen E, Sakari-Rantala R, Rantanen T (2007) Fall-related injuries among initially 75- and 80-year old people during a 10-year follow-up. Arch Gerontol Geriatr 45(2):207–215

Scottish Intercollegiate Guidelines Network (2015). Critical appraisal: notes and checklists. Retrieved from http://www.sign.ac.uk/ methodology/checklists.html

Smith RW, Nelson DR (1998) Fractures and other injuries from falls after an ice storm. Am J Emerg Med 16(1):52–55

Stalenhoef PA, Crebolder HF, Knottnerus JA, Van Der Horst FG (1997) Incidence, risk factors, and consequences of falls among elderly subjects living in the community: a criteria-based analysis. Eur J Pubic Health 7(3):328–334

Sterling DA, O'Connor JA, Bonadies J (2001) Geriatric falls: injury se- verity is high and disproportionate to mechanism. J Trauma 50(1): 116–119

Stevens JA, Thomas KE, Sogolow ED (2007) Seasonal patterns of fatal and nonfatal falls among older adults in the U.S. Accid Anal Prev 39(6):1239–1244

Todd C, Skelton D (2004) What are the main risk factors for falls among older people and what are the most effective interventions to prevent these falls? WHO Regional Office for Europe, Health Evidence Network report, Copenhagen

Turner RM, Hayen A, Dunsmuir WT, Finch CF (2011) Air temperature and the incidence of fall-related hip fracture hospitalisations in older people. Osteoporos Int 22(4):1183–1189

Vikman I, Nordlund A, Naslund A, Nyberg L (2011) Incidence and sea- sonality of falls amongst old people receiving home help services in a municipality in northern Sweden. Int J Circumpolar Health 70(2): 195–204

Yeung PY, Chau PH, Woo J, Yim WT, Rainer TH (2011) Higher inci- dence of falls in winter among older people in Hong Kong. J Clin Gerontol Geriatr 2(1):13–16

Zautcke JL, Coker SB Jr, Morris RW, Stein-Spencer L (2002) Geriatric trauma in the state of Illinois: substance use and injury patterns. Am J Emerg Med 20(1):14–17

2088 Int J Biometeorol (2018) 62:2073–2088

Reproduced with permission of copyright owner. Further reproduction prohibited without permission.