22 Pages
Gain access to the library to view online
Learn more

Aides aux personnes âgées dépendantes : La famille intervient plus que les professionnels (version anglaise)


Gain access to the library to view online
Learn more
22 Pages


Une personne de plus de 65 ans sur dix vivant en domicile ordinaire ne peut sortir sans aide. À incapacité comparable, les aides dont bénéficient les personnes dépendantes restent très inégales : si 55 % des personnes sévèrement dépendantes reçoivent plus de trois heures d'aide par jour, 20 % en reçoivent moins d'une heure. Certes, cette variabilité traduit, en partie, la difficile quantification des aides informelles fournies par les proches. Elle montre, cependant, qu'il serait délicat d'établir un barème associant un volume d'aide à un niveau de dépendance. La durée d'aide totale d'aide augmente avec la dépendance et l'isolement, mais ne semble pas liée au niveau d'éducation ou au revenu. Cependant, les facteurs économiques influencent fortement le prix horaire moyen et donc la dépense totale : les personnes les plus aisées ne bénéficient pas, à autres caractéristiques comparables, d'aides plus importantes mais les paient plus cher. L'aide informelle reste beaucoup plus développée que l'aide professionnelle : à tous niveaux de dépendance, la moitié des personnes âgées ne reçoit que ce type d'aide. Aux niveaux de dépendance modérée, le temps d'aide informelle représente 2 à 3 fois celui d'aide professionnelle. L'isolement, le niveau d'éducation et une résidence dans l'Ouest, le Centre-Ouest ou le Bassin parisien favorisent le recours à l'aide formelle. Le temps d'aide formelle reçu est alors d'autant plus important que son prix horaire moyen est faible. Il est également plus élevé lorsque les dépenses associées ouvrent droit à des réductions d'impôts. Toutefois, les volumes d'aides formelle et informelle sont trop variables d'une personne dépendante à l'autre, pour que l'on puisse déterminer dans quelle mesure ces deux types d'aides peuvent se substituer.



Published by
Reads 48
Language English


Caring for the Dependent
Elderly: More Informal
than Formal*
One in ten people over 65 years old and living at home cannot get out without
help. For comparable disability levels, there are still major imbalances in carePascale
given to dependent individuals. Whereas 55% of highly dependent individualsBreuil Genier**
receive more than three hours of care per day, 20% receive less than one hour.
Although this disparity partially reflects the problems with quantifying informal
care from the family, it also shows that it would be hard to establish a scale
associating volumes of care with levels of dependency.
Total hours of care increase with dependency and isolation, but do not appear to
be linked to the level of education or income. However, economic factors strongly
influence the average price per hour and thus total expenditure. Other
characteristics being comparable, the wealthiest individuals do not receive more
care, but pay more for it.
Informal care is still a lot more developed than formal care. For all levels of
dependency, half of the elderly only receive informal care. For moderate levels of, two to three times more informal care than formal care is received in
terms of time. Isolation, level of education and residence in the West, Central West
and the Paris basin favour receiving formal care. The lower the average hourly
price of formal care, the longer the hours received. Hours of care also lengthen
when the associated outlays are tax deductible. Nevertheless, formal and informal
care volumes vary too much from one dependent person to the next to be able to
determine the extent to which these two types of care are substitutable.
* Originally published as
“Aides aux personnes
âgées dépendantes : la
famille intervient plus
que les professionnels,”
Économie et Statistique,
no. 316 317, 1998. n France, there are still fairly few statistics grids have been developed (see box 1). Above
** Pascale Breuil Genier I available on the dependency of senior and beyond the problem of defining a situationis head of the Health
Economics Bureau in citizens (Join Lambert et al , 1997). of dependency is the question of how to
the Social Security Dependency and the needs it creates are hard to respond to it. There is no easy standard way of
quantify. First of all, the very notion of dealing with it. The attractive idea of a standard
The names and dates in
dependency is quite vague. Dependency, i.e. type of care for each level of dependency is
brackets refer to the
the need for assistance from a third person, is borne out neither by Joël and a Martin’sreferences at the end of
the article. subjective notion for which many evaluationsociological analyses (1996) nor by the
INSEE Studies, no. 39, September 1999 1Box 1
Objective and subjective definitions This research posits that two main elements can
be used to identify disadvantaged individuals:
The international classification of impairments, being homebound and dependency on a third party
disabilities and handicaps developed by the World to wash and dress. The 8,350,000 individuals over
Health Organisation helps clarify the concepts 65 years old living at home (94%) and in
generally associated with the notion of institutions (6%) are divided into four categories
dependency (CTNERHI, 1988): using the Colvez classification: bedridden and
chairridden (230,000) (level 1), not bed or
– Impairment denotes any loss or abnormality of a chairridden but helped to wash and dress
psychological, physiological or anatomical struc- (320,000) (level 2), neither bed/chairridden nor
ture or function. It is evaluated with reference to a helped to wash and dress but helped to get out of
biomedical norm and cannot be perceived by the the house (940,000) (level 3), independent (level
impaired individual. 4) (Kerjosse and Lebeaupin, 1993). Three
additional categories (bearings in time and space,
– Disability corresponds to a restriction or lack (re behaviour, and help with bearings and behaviour)
sulting from an impairment) of ability to perform are used to refine this classification by dividing
any activity in the manner or within the range con each of the first four categories into two
sidered normal for a human being such as getting sub categories according to level of psychological
up, getting around, etc. The important factor here dependency.
is not the deviation from an ideal (and theoretical)
state, but solely the objective manifestations of The Isoresources Group Gerontological
this deviation. Independence grid (AGGIR or GIR) is designed to
evaluate the work load associated with the
– Handicap denotes a disadvantage for a given indi- dependency of seniors living at home and in
vidual, resulting from an impairment or disability, institutions. It is based on ten discriminating
that limits or prevents the fulfillment of a role that elements concerning coherence, bearings,
is normal for that individual (according to the cultu washing the top and bottom half of the body,
ral norms of their environment). For example, dressing (top, middle and bottom), eating, urinary
individuals who are relatively homebound (disabili and faecal elimination, transfers, getting around
ty) could consequently find their social network indoors and outdoors, and communicating over
reduced (handicap). distances. Depending on what the senior does,
does not do or partially does in these ten areas, he
Dependency in the strict sense of the term is or she is classed by a complex algorithm in one of
therefore defined as the need for a third party in the six isoresources groups representative of a
activities of daily living. Along with the diversity of mobilisation of caregiver means. The six groups
the data used, the many different concepts have obtained can be described in outline as follows.
produced a huge number of measurement scales. Group 1 denotes individuals who have lost their
Users each need a measurement perfectly suited mental, bodily, motor and social independence and
to the use that they want to make of it, even if thisabsolutely need caregivers to be present all the
undermines the comparability of the results. time. Group 2 is made up mainly of “lucid invalids”
and the “ambulant demented”. Group 3 generally
Assessing the needs of the elderly covers individuals who still have their mental
independence and partially their motor
The purpose of the dependency measurement independence, but who need help with their bodily
obtained from Kuntzmann’s score is to assess the independence several times a day. Group 4 covers
human resources requirements of the individuals individuals who need to be helped or stimulated to
studied. Nine indicators describe five fields (human wash and dress and cannot get up alone but, once
resources requirements, getting around, up, can get around their dwelling. It also covers
continence, psychological state and care needs). individuals who have no motor problems, but need
These are used to ascribe a dependency score of help for bodily activities including eating. The
0 to 10. This method was applied to the 1987 1988individuals in group 5 need ad hoc care (usually
CREDES Seniors in Institutions survey (Mizrahi home help) and those in group 6 are considered to
and Sermet, 1989). be independent (Vetel, 1994a and b).
The French Ministry of Social Affairs’ Department The definitions used by private insurers vary from
of Statistics, Studies and Information Systems one contract to the next. Most of the contracts
(SESI) also has a dependency evaluation grid require the permanent impo ssibility of
(Kerjosse, 1992). Five categories of ph ysical independently carrying out one of the activities of
dependency (eating, washing, dre ssing, daily living (washing, getting around, eating and
incontinence and getting around) are used to class dressing). Some contracts are also based on
the elderly into four groups. INSERM research social security payments (lump sum payments for
helps define these physical dependency groups. home or institutional care). Others include the
2 INSEE Studies, no. 39, September 1999statistical analyses presented in this article.
Box 1 (end) Care needs are highly diverse (care, help with
activities of daily living, special
accommodation, moral support, supervision,
longest length of independence in the day or etc.). A number of combinations of care could
the cumulative length of care over 24 hours. satisfy them. The optimal care plan would
The 1996 Household Living Conditions panel
therefore seem to be closely linked to the
survey on local services included a variable
preferences of the senior citizen and his or hermeasuring the maximum length of time that the
person could stay alone. This question was family. Lastly, the usual statistical surveys are
used in particular to identify psychological not really suited to collecting data on
dependency. dependency and consequently do not provide
much information on this subject. On the one
Estimating the state of health
hand, dependency only affects a small
The CREDES eight class invalidity indicator proportion of the elderly population (Kerjosse
provides objective information on the population and Lebeaupin, 1993). This means using a huge
studied. The consequences in terms of the survey sample and a filtering questionnaire
disability of ailments, impairments and where
to collect data on a sufficient number of
appropriate the treatment of each interviewee
dependent people. On the other hand,are evaluated by a doctor based on medical
knowledge at the time of the measurement. traditional surveys are generally of ordinary
This indicator was calculated using the households whereas many dependent senior
1991 1992 INSEE Health survey of seniors citizens live in institutions (long term care
living at home and the 1987 1988 CREDES
centres, old people’s homes, etc.).
survey of seniors over 80 years old living in
institutions. The Health survey found that 7.5%
of the over 65s and 15% of the over 80s had a
very limited domestic activity in 1991 (indicator Few surveys are available on care given
classes 5 to 8). This figure rose to 78% for the to the dependent elderly
over 80s living in institutions (Mizrahi and
Mizrahi, 1994).
1 The surveys referred to often cover limited
The central criterion in the calculation of geographic areas and are sometimes a little old.
disability free life expectancy is mobility One of the most well known is the 1988 survey
(Mormiche and Robine, 1993 and 1994). French national pension fund of the two French
Disability is studied here over the entire life
provinces of Doubs and Loire Atlantique. This
cycle and is therefore not based on criteria
survey covered 2,000 people over 75 years olddevised for the elderly. For those living at
home, the short term disabilities (retirement) living at home or in an institution. Information
combine with severe long term disab ilities was collected on their level of dependency, the
(cannot get up or can get up a little) and professional and voluntary help they received
moderate long term disabilities (can go out with
and their health care expenditure (obtained
the help of another person, an assistive device
directly from the health insurance funds). Theor walking stick or needs no help, but has
certain limitations or difficulties). For those survey found that only 5% to 10% of extremely
living in institutions, severe disability takes the dependent individuals living at home received
form of being bedridden or chairridden or solely professional care, while over 50% only
needing another person to wash and dress,
received care from the family and 40%
while moderate disability corresponds to help to
received both types of care.leave the estab lishment or psychological
dependency. The 1991 1992 INSEE Health
survey data and the data provided by the 1990 Another frequently cited source of data on
SESI survey of establishments caring for the care to the dependent elderly was put together
elderly (EHPA) found a disab ility free life
by Colvez et al. (1990) using surveys carried
expectancy of 63.9 years for men (out of a total
out by three regional health observatorieslife expectancy of 72.9 years in 1991) and a
disability free life expectancy of 68.8 years for (Basse-Normandie, the Paris area and
women (out of a total life expectancy of 81.1 Languedoc Roussillon). These surveys
years). Life expectancies free from severe covered 5,000 people aged over 65 years living
disability are 71.9 years for men and 79.3 years
at home or in an institution. Of the 2.4%
for women.
bedridden and chairridden individuals
(one-quarter of whom lived in institutions),
47% received no formal care (not even a private
For example, in the Commissariat général du Plan report
INSEE Studies, no. 39, September 1999 3cleaning lady) and 34% received both elderly into groups defined by the “Isoresources
professional and family care. Group Gerontological Independence” grid
(AGGIR or GIR; see box 1). This grid is used
Last but not least, the 1992 CNAV mainly to define entitlement to attendance
2Intergenerational Relations and Family benefit.
Support survey of 2,000 adults aged 49 to 53
and their parents found that 50% of the elderly
Dependency levels parents received regular care from a member of
their family. This percentage rose to 84% when are also linked to levels
the elderly person found it difficult to carry outof education
activities of daily living. The total number of
services provided to the elderly does not appear The probability of being dependent varies
to depend on the income of the caregivers according to the senior’s characteristics.
(Attias Donfut, 1996). Econometric analyses reasoning on the basis of
“other things being equal” show that age is the
This article presents new findings on care to theprime explanatory factor in the dependency of
dependent elderly and discusses the conceptualseniors living at home. Gender is found to have
and methodological problems affecting their no influence, possibly because dependent
interpretation. It draws on data from INSEE’swomen often live in institutions.
1996 Household Living Conditions panel
survey of local services. This national surveyOther things being equal, a low level of
interviewed 2,211 people aged over 65 and education increases the probability of being
living at home. It was combined with two local dependent. Although no statistical link has
extensions (Paris area and Eure et Loir) to been found between income and dependency,
bring the number of elderly people interviewednumerous standard of living indicators have
up to 3,520 (see box 2). proved to be linked to dependency. For
example, people living in a “wealthy” or
“financially sound” household are less likely to
10% of seniors living at home be dependent than those who have “trouble
are slightly dependent making ends meet” or “cannot manage without
credit” (9.5% of the elderly). The level of
A number of variables can be used to describe dependency does not appear to be linked to the
the dependency levels of the elderly people main housing sanitation factors such as no
studied (see box 1). The first, when slightlyindoor toilet (4% of seniors), no bathroom
adjusted, was used as a filtering variable in the (6%), no hot running water (4%) and no central
1996 Household Living Conditions panel or electric heating (16%). Neither does living in
survey and is based on the classification drawn a private flat (29%) or house have any influence
up by Dr Colvez. The dependent elderly areon dependency. These findings differ from
defined as those who are bedridden or those of Colvezet al . (1990), who found that,
chairridden (2.4% in the survey), those who are depending on the levels of dependency, 12% to
helped to wash and dress (a further 3%) and18% of dependent seniors had no hot running
those who cannot go out without assistance water, 10% to 18% had an outdoor toilet and
from a third party (a further 4.3%). In total,8% to 16% complained of inadequate heating.
10% of the over-65s living at home are These percentages are higher than those found
dependent according to this definition. Thesefor all seniors and suggest that dependent
people and those who said they received individuals have fewer modern conveniences.
dependency care were asked to give their own
opinion of their state of health. To round outOther senior characteristics also appear to be
these first two variables of Colvez defined correlated with the level of dependency.
dependency and subjective state of health, However, these were not included in the final
which rather describe physical dependency, the
individuals who said they received dependency
care were asked to state the maximum amount
of time they could stay on their own. This was
done to get a better idea of situations of
psychological dependency, which create
2 Attendance benefit was introduced by the law of 24 January
substantial care needs. Lastly, the Eure et Loir1997. It is a benefit in kind available only to persons in
extension collected information used to sort the isoresources groups 1 to 3 (decree of 28 April 1997).
4 INSEE Studies, no. 39, September 1999model either because they were too closely What is the definition
correlated with a characteristic already taken
of dependency care?
into account (such as standard of living
indicators), too close to the variable to be
explained (such as health indicators) or his last finding suggests that it would be
because they might have been a consequenceT preferable in a study of the choices of
rather than a cause of dependency (such as dependent seniors to analyse both the choice
household size). For example, other financial between professional and family care for those
situation indicators were also able to predict living at home and the choice being staying at
the level of dependency: seniors in a home and going into an institution. Gramain
household finding it hard to meet current (1997) defines a number of configurations: just
expenses in the previous two years (4%) and formal care at home, the use of formal care at
those who had to defer payment of their rehome, in an institution, etc. nt However,
(1%) were found to be more dependent. individuals living in institutions are not
Conversely, being a home owner or a tenancot vered by the 1996 Household Living
is not indicative of the level of dependencyConditions. panel survey on local services.
Moreover, for given sociodemographic Therefore, only the choice between formal and
characteristics, health problems sharply informal care at home is analysed here. The rest
increase the probability of dependency. of the study concentrates just on the sample of
Seniors treated regularly for a serious or 685 individuals who said they received at least
chronic illness, those with a poor state of one type of dependency care (see box 2). This
health and those who have had to give up has been done to better isolate the elements
work due to health reasons in the last twelve likely to influence this choice of elements
months are more likely to be dependent. determining the need for care.
Likewise, health care consumption is a good
gauge of dependency. Dependency is found Yet even for these 685 individuals, defining
more frequently among the elderly who havedependency care is not as easy as it seems.
been hospital in patients in the last twelveWhere does the natural mutual care by
months, those who have seen a doctor severalmembers of the same household end and
times and those who regularly take dependency care start? Should a cleaning lady
medication. Individuals who worked in the be considered to be dependency care? This
past are less likely to be dependent. This study counts all the assistance with everyday
would appear to reflect a protective effect of life declared by a household with a dependent
work rather than a selection effect. A senior as being care received by this person (see
selection effect would mean that individualsbox 2). This choice is all the more justified
with a severe disability at a fairly young age since the assistance declared as home help and
would have less access to employment not as dependency care, which could possibly
and a greater probability of still being be considered to be convenience assistance,
dependent after 65 years old. Yet the was generally received for the first time at the
influence of past employment remains start of the senior’s dependency.
significant when the sample is restricted to
individuals who say they became dependentEach type of assistance, whether to the
after the age of 65. household or one of its dependent members, is
described separately: type of assistance giver,
Lastly, the type of household in which the frequency and hours, and remuneration. For
senior citizen lives appears to be closely linkedhousehold assistance, nature (housework,
to dependency. The least dependent are shopping, cooking, etc.) and how long it has
individuals living alone, followed by those been provided are specified. A volume of hours
living with a partner. However, those who live of assistance per week should, in principle, be
with their children or another family memberable to be calculated from these elements.
are more frequently dependent. The influenceHowever, it is very hard to quantify informal
of household size therefore probably points to a care provided by a member of the same
selection effect. When people living alone household, especially for highly dependent
become dependent, they no doubt have to goindividuals. Should the actual care time or the
into an institution or go to live with family.hours of presence when the dependent person
Household size therefore appears to be more of needs constant supervision be counted? It is
a consequence than an explanatory factor ofalso harder to quantify assistance to a
dependency. household member than to evaluate the length
INSEE Studies, no. 39, September 1999 5Box 2
Questionnaires and samples Study of the probability
of being dependen t
The national questionnaire The study of the probab ility of being dependent
The data used in this study are taken from the Maycan be approached by analysing the dependency
1996 Household Living Conditions panel survey measured by the Colvez classification or directly
with a variable section on local services. This by the existence of assistance in activities of daily
INSEE survey contains three sections: living. The study of disability measured by the
A questionnaire on the household’s usual Colvez classification can be made for severe
sociodemographic information (Q1), or moderate disability (Colvez levels 1 to 3)
A questionnaire on health, housing and concerning 197 of the 2,211 individuals in the
indebtedness (Q2), national sample or approximately 10% of this
A questionnaire on local services (Q3). sample after weighting (see appendix). The study
of recipients of assistance in activities of daily
In the part of the study on local services (Q3), living only differs from the previous study in that it
households are asked, among other questions, uses adjusted Colvez levels 1 to 4 to define
about the formal and informal care they receive dependency and is based on the assumption that
(excluding dependency care) (section Q3.1 of the this variable properly identifies individuals
questionnaire). Dependency care data is collected assisted with activities of daily living (by excluding
by a Kish questionnaire, that is a questionnaire put care and services more for reasons of
to an individual sampled at random from among convenience than disability).
the dependent members of the household (section
Q3.2 of the questionnaire). These dependent The notion of dependency care
individuals are filtered using an adjusted Colvez The survey asks respondents not to declare
classification: level 4 was added to collect dependency care in the section of the
information on the care received by slightly questionnaire reserved for assistance received
dependent individuals, which brought the total by the household (Q3.1). Information on
number of levels to five: dependency care is collected in section Q3.2 of
Level 1: the individual cannot get out of bed or a the questionnaire. The question is put here as to
chair without help from a third party, whether the dependent person receives other
Level 2: the individual needs help washing and help aside from the paid help mentioned in
dressing, section Q3.1. The formulation of this question,
Level 3: the individual cannot leave home without like the difficulty of effectively distinguishing
help from a third party, between dependency care and traditional home
Level 4: the individual is a ssisted in activities of help for a dependent person, means that
daily living (shopping, housework, etc.), assistance of the same nature can subsequently
1Level 5: the individual receives no care . no longer be distinguished according to the
questionnaire section from which they were
2The national survey collected information on taken . More precisely, the following rule was
14,845 individuals (2,211 of whom were elderly) adopted. Any care declared in Q3.1 concerning
living at home. housework, shopping or cooking dating no more
than two years before the first dependency care
The geographic extensions (if this exists) is added to the help declared in
3 Two extensions were added to this survey. The Paris Q3.2. This rule does not consider help with
area extension used the same methodology to collect gardening, laundry, window cleaning, sewing or
information on 2,479 individuals (251 of whom were administrative tasks to be dependency care.
elderly). The Eure-et Loir extension was designed to There are a number of reasons for the choices
study care received by seniors and was based on an made. The first is to make dependency care
abridged questionnaire (minus the Q1 and Q2 survey more homogeneous and focused on essential
sections), which was slightly modified (the questions daily activities (by excluding activities assumed
on dependency care were put to one elderly person to be characteristic of convenience care).
per household without any condition of dependency). Secondly, this strategy prevents a person using a
This second extension collected information from gardener three times a year from being deemed
1,058 aged individuals. Specific use was made of it in a formal care recipient. More generally, it
Robert (1997). The national survey and the two prevents a reduced volume of care from having
extensions were combined to form a preliminary an influence on the receipt of formal and informal
sample of 3,520 seniors, 1,509 of whom answered the care. Lastly, the exclusion of care too far in the
Kish questionnaire (see above) on dependency care. past or of a particular nature only has a marginal
effect. Excluding all the assistance declared in
Methodology Q3.1 would only reduce the probab ilities of
A number of different types of studies can be made receiving formal and informal care by 2% to 4%.
using such a data base:
6 INSEE Studies, no. 39, September 1999Box 2 (end)
The division between formal and informal care individuals (i.e. those who answered questionnaire
Two strategies can be used to study the d ivision Q3.2 see above) who say they receive at least
between formal and informal a ssistance in one type of dependency care. This population
activities daily living: totals 685 individuals and its structure is presented
- Either look at the probabilities of receiving each in the table.
type of assistance for the individuals who declare
help in activities of daily living (adjusted Colvez Throughout the study, we have tried to see
classification level different to 5 and existence of whether the correction of any selection biases
4care); could modify the influence of economic
- Or look at the probabilities of receiving each type determinants on the use of care in the event of
of care for the entire aged population. dependency. Two types of selection effects were
The second strategy has the advantage of using a studied: selection effects linked to being
larger sample, but the disadvantage of not dependent and, for the analysis of formal and
distinguishing the determinants of the need for informal care, selection effects associated with
care and those of the choice between different using formal or informal care. To do this, we
types of care. The choice between the two introduced Mills ratios into the equations
strategies also depends on the underlying determining the volumes of care received (or
behavioural model. The first infers that the fact that expenditure or prices). These ratios were never
individuals will receive care is determined first (by found to be significant and never distinctly
a behavioural model or, more in line with this changed the influence of the other determinants
strategy, by objective elements of need) and that introduced into the equation.
the individuals then decide between formal and
informal care. The second strategy is more suited
to a behavioural model in which individuals make
1 The last level of this variable concerns whether care istwo decisions using formal care and using
received and not whether there is a need for care. This is justifiedinformal care possibly simultaneously, but without
by the fact that the aim of this questionnaire is to study actual
their choosing process necessarily leading them to
care use behaviour: this filtering ensures, at least in theory, that
ask for at least one of the two forms of care.
only those individuals who receive care are interviewed.
Choosing the first approach is tantamount to 2 This seems to be all the more justified in that over half of the
forming a hypothe sis on the magnitude and people making up our sample of those living alone identify home
objectiveness of the dependency needs and the help with personal care.
3value of limiting the study to the most We also check that the care is not declared in both parts of
the questionnaire. When this is the case, one only of the twohomogeneous population possible from the point
is retained (for 50 individuals).of view of its motivations (help to counteract a
4 This second condition is to exclude the 28 level 4 individualsdisability rather than for convenience).
who said they received no help, which we consider to be
erroneous data.
We favour the first approach (see appendix). The
study sample is therefore made up of Kish
Structure of the aged population living at home and receiving dependency care
Number and % of individuals
Ability to remain alone State of health (given your age,
Adjusted Colvez ( the longhest amount of time how would you Isoresources group
classification the person can stay alone view your current (GIR)
without care) state of health)
Level 1: bedridden Less than one hour Very poor Group 2
or chair ridden
74 (11%) 73 (11%) 94 (14%) 15 (7%)
Level 2: helped Less of half a day Fairly poor Group 3
with washing and dressing
90 (13%) 159 (23%) 353 (52%) 24 (11%)
Level 3: cannot leave home, One day Normal Group 4
without help from a third party 126 (18%) 181 (26%) 28 (12%)
146 (21%)
Less than a week Fairly good Group 5
144 (21%) 51 (7%) 39 (17%)
Level 4: Receive other One week or more Excellent Group 6
375 (55%) 182 (27%) 5 (1%) 118 (53%)
Total numbers
685 (100%) 684 (100%) 684 (100%) 224 (100%)
(1 missing value) (1 missing value) (Eure et Loir sample)
Coverage: aged individuals living at home who says they receive dependency care, i. e. those who answered the Kish questionnair e put
to individuals randomly sampled from among the dependent members of the household (sectionQ3.2 of the questionnaire).
Source: 1996 INSEE Household Living onditions panel surveyC .
INSEE Studies, no. 39, September 1999 7of visits to a family member’s home. Therefore, Yet these hours of care are highly variable.
the analyses of care volumes in this study use a Some 20% of bedridden and chairridden
“time brackets” approach rather than an individuals and those helped to wash and dress
average, since the latter could be too sensitive and 15% of individuals who say they cannot
to the quantification of extreme volumes of care. stay alone for more than one hour receive less
than one hour of care per day. Conversely, 21%
of individuals who only need help to get out of
30% of dependent the house state that they receive more than three
individuals receive hours of help per day.
over three hours of help per day
Extremely variable hours of care oftenThe volume of care received by senior citizens
who say they receive dependency care is below theoretical needs
closely linked to their level of disability (see
table 1). If the adjusted Colvez classification is The median volume of care for a dependent
used as a disability indicator, 55% of extremely person is seven hours a week. It rises to nearly
dependent individuals (level 1 and 2: bedridden50 hours for bedridden and chairridden
or chairridden or helped to wash and dress) individuals. These care norms drawn up by
receive over three hours of care per day as experts using a list of needs can be compared to
opposed to 21% on average for other dependent find whether the care received by dependent
individuals (level 3 and 4). Similarly, 58% of seniors corresponds to their needs. Colvez et al .
individuals in isoresources group (GIR) 2 and 3 (1990) endeavoured to establish such a norm by
(heavily dependent) receive over three hours ofasking a dozen experts to calculate the time
care per day as opposed to 33% of GIR 4 and 5 required to provide certain assistance in
(moderately dependent) and 20% of GIR 6 activities of daily living to dependent seniors.
(independent, see box 1). Logically, the totalAnswers varied a great deal, which shows how
volume of care seems even more closely linked hard it is to theoretically evaluate the hours of
to the maximum amount of time that care required by a dependent person. The
individuals say they can stay alone: 67% ofexperts reckoned, for example, that
those who say they cannot be alone for moreone and a half to three hours a day were
than one hour receive over three hours of care necessary to feed a dependent individual (the
per day, compared with only half of those who median being two hours). Evaluations for
can be alone for half a day and 5% of those who washing also varied from a half an hour to one
can be alone for over a week. hour with a median of 35 minutes. Opinions
Table 1
Breakdown of the total volume of care
In %
Dependancy indicators
Hours of care Adjusted Colvez (1) Isoresources group (2) Maximum length of independance (alone)
per week
Levels Levels GIR GIR Half Under Over
GIR 6 1 hour Day
1 and 2 3 and 4 2 and 3 4 and 5 a day one week one week
Less than 1h30 9 3 10 3 1393397 15
From 1h30 to 3 h 18 5 22 5 7 33462 26 38
From 3 h to 7 h 23 12 271322291015243327
From 7o 21 21 25 2021251728262015
Over 21 h 29 55 2158332067503514 5
Total 100 100 100 100 100 100 100 100 100 100 100
1. Adjusted Colvez classification: levels 1 and 2 = extreme dependency; levels 3 and 4 = moderate or little dependency.
2. Isoresources group(GIR) : the higher the GIR level, the lower the dependency (see box1).
Coverage: aged individuals living at home who say they receive dependency care.
Source: 1996 INSEE Household Living Conditions panel survey.
8 INSEE Studies, no. 39, September 1999diverged even more as to the time needed to model the breakdown by care time intervals
prepare meals (from half an hour to four hours (ordered logit).
with a median of one and a half hours).
Dressing was estimated at between a quarter of
Dependency level and living an hour and two hours (median at 40 minutes),
and so on. Based on the above median valuewith os, ther people determine
assistance to wash, dress and prepare meals total care time
would take approximately 20 hours a week. Yet
20% of extremely dependent individuals The level of dependency is obviously the main
(Colvez classification 1 and 2) say that theydeterminant of the total volume of care. The
receive less than seven hours of care per week. maximum amount of time a person can stay
This finding raises a question about the living alone has greater explanatory power than the
conditions of dependent persons living at disability measured by the Colvez
home. However, all of these figures should be classification. The isoresources group has less
6viewed with caution, since neither the explanatory power and the subjective state
quantification of care by the interviewees norof health has none. For a given dependency
the norms defined by the experts can be level, the hours of care no longer vary with age
completely trusted. For example, if the surveyand sex.
results are to be believed, over three quarters of
the individuals interviewed receive all the careThe second determinant of the amount of care
3 3 Only 130 people they need. Where a lack of care was time is the aged person’s household structure.
answered this question, mentioned, 58% mentioned housework and The elderly living with a child (14% of all
which should have been
gardening, 40% shopping, 40% getting to seniors and 12% of dependent individuals)put to everyone, but was
asked following a appointments (doctors and others), 36% receive the most hours of care, followed by
complicated filtering someone to talk to, 22% administrative tasks,those who live with other family members
18% feeling safer at home, 14% dressing and (12% and 16%). Those who live with a partnerConsequently, and also
because very small 11% taking part in a leisure activity. These are (44% and 27%) and those who live alone (30%
numbers are concerned, highly fragile estimates and must be viewedand 45%) are the least assisted. Household
the findings should be
with great caution. Moreover, the individualsstructure measures the “potential pool ofviewed with even more
caution. who said they lacked care did not appear to family care”, but can also be endogenous with
receive less care than the others (see infra). the most dependent individuals being more
likely to live with other family members.
These findings show that it is virtually impossible
to associate a standard number of hours with a Receiving both formal and informal care results
level of disability. This justifies the decision in a higher volume of care. This variable may
concerning the recently introduced attendance either indirectly measure the dependency level or
benefit to call on a medico social team to evaluate reflect the fact that care shared among a number
4 4 This team goes to the needs of the dependent person. of caregivers means that each can give more.
benefit applicants’ homes
and, if appropriate, meets
Since, even for comparable dependency levels, Lastly, a greater amount of care was reportedtheir families to assess
the standard of facilities the amount of care received by dependent when one of the caregivers answered the survey
in their dwelling and the persons is highly variable, can part of the instead of the elderly person. Care recipients
assistance they receive
variability be explained by objective elements and caregivers can have a different view of carefrom their next of kin.
such as family structure and income? An volumes. Frossard et al. (1988) noted that
econometric approach reasoning on the basis of dependent individuals reported an average of
“other things being equal” is well suited to this 76 hours of care compared with 90 hours
question. It separates out the influences of the reported by their caregivers. However, this
different characteristics of the dependent effect could also be endogenous. Firstly, the
5 Econometricians may 5individuals and their next of kin. more care given by a caregiver, the more likely
be disappointed with the
he or she is to be present when the survey ismodels estimated since
few variables are The study’s sample for care volumes is made up
significant. Nevertheless, of all the elderly who report at least one type of
this result could be
dependency care and for whom a length of timeinterpreted positively by
saying that few nequalities i was able to be reconstituted (totaling 648
associated with objective, 6observations, see box 2). Given the The isoresources group is only significant at the 5% level for
known characteristics are the sample of dependent individuals in the Eure-et-Loiruncertainties in the quantification of theobserved in care for a extension as opposed to 0.01% for the same sample with regard
given level of disability. number of hours of care, particularly long to disability (Colvez) and the maximum amount of time a person
hours of care, non linear models were used t can be alone.o
INSEE Studies, no. 39, September 1999 9taken. Secondly, since the priority interviewee isWhat is the breakdown
the elderly person, the caregiver only answers
between formal and informal care?
when the former is absent or too dependent to
answer. No other indicator of informal care proved
to be significant. In particular, the proximity and lthough 72% of the French population
availability of family members plays no part. A considers that children should look after their
Could it be that dependent individuals and theiraged parents in need, 49% are in favour of
children live very close to one another? The 1992 dependency care provided essentially by the next
CNAV survey found that 49% of parents with a of kin (as opposed to 28% in favour of state and
child aged 49 to 53 years live less than one local community services and 22% in favour of
kilometre from this child and 90% live less than s 50ocial security and supplementary insurance
km away (Attias Donfut, 1996). schemes) (Credoc, 1995). Approximately 80% of
extremely dependent individuals (bedridden,
However, for a given level of disability, total chairridden or needing help to wash and dress)
care time does not appear to be linked to the receive care from their family, with approximately
dependent person’s socioeconomic characteristics 50% of them receiving solely this type of care. Just
(income, level of education and socio professionalunder half receive professional care, with one fifth
7 7 Admittedly, if the group). of highly dependent individuals receiving no other
average price paid by the type of care. One third receive both formal and
dependent person per
Therefore, the total volume of care received informal care (see table 2). These proportions are inhour of care is introduced,
it has a highly significant by dependent seniors is essentially a function the main comparable with those found by previous
negative impact on the of their dependency level and the structure ofsurveys except that they indicate that a higher
length of care. However,
the household in which they live (see charts proportion of highly I dependent individualsthis effect cannot be
8interpreted as a price and II). Yet a hypothesis could be advanced receives just formal care.
effect, since the average that an aged person’s socioeconomic
price calculated here also
characteristics influence more the type ofreflects the proportion of
informal care (low cost) in care received than the total volume of care8. The increase in dependency care since the late 1980s could
total care and the Separate analyses of formal and informal partially explain this development. An analysis of the attendance
individuals who receive
benefit test in 1995 showed that 34% of the care plans definedcare followed by analyses of the elementsmainly informal care
by the medico social teams recommended family care (20% of
generally receive higher involved in the choice between these two which exclusively) and 58% recommended home help services
volumes of care. forms of care confirm this. (34% of which exclusively) (CNAV, 1996).
Chart I
The greater the disability, the more care received
1. Colvez grid: levels 1 and 2 = extreme dependency; level 3 = moderate or little dependency.
2. The higher the GIR level, the lower the dependency.
See box 1 on dependency grids.
Coverage: aged individuals living at home who say they receive dependency care.
Source: 1996 INSEE Household Living Conditions panel survey.
10 INSEE Studies, no. 39, September 1999