Full text of DOI:10.1093/geront/41.5.589

The Gerontologist
Vol. 41, No. 5, 589–596
Copyright 2001 by The Gerontological Society of America
Resident Contact With Family and Friends
Following Nursing Home Admission
Cynthia L. Port, PhD,1 Ann L. Gruber-Baldini, PhD,1 Lynda Burton, ScD,1
Mona Baumgarten, PhD,1 J. Richard Hebel, PhD,1 Sheryl Itkin Zimmerman, PhD,2
and Jay Magaziner, PhD, MSHyg1
Purpose: This study explored factors that are related to
the level of contact (number of visits and calls) between
newly admitted nursing home residents and their family
and friends. In addition to reexamining factors studied
previously, several new factors were explored: contact
level prior to nursing home placement, dementia status,
and resident race. Design and Methods: Interviews
were conducted with the significant others of 1,441 resi-
dents from a representative sample of nursing homes
in Maryland. Results: Contact decreased by approxi-
mately half following admission, compared to reported
preadmission contact. Rates of contact are positively re-
lated to nonuse of Medicaid, kinship closeness, support
network proximity, nondemented status, and White race.
After controlling for preadmission contact, postadmission
contact is positively associated with kinship closeness,
support network proximity, nondemented status, and
White race. Implications: The study identifies factors
that are useful to consider when designing interventions to
increase family involvement with nursing home residents.
derstanding is unfortunate because family involve-
ment has the potential to positively influence the
nursing home setting through improved psychosocial
well-being of residents (Greene & Monahan, 1982),
reduced care burden of staff, and increased public in-
volvement in the concerns of America’s enlarging el-
derly population. With better information about fac-
tors that inhibit or facilitate involvement, interventions
may be feasible.
Part of the reason for the lapse in research focus
may be the longstanding misconception that nursing
home residents are abandoned by relatives. In fact,
family members and friends stay in contact, retain
close emotional ties, and continue to contribute to
basic care needs of residents (see Naleppa, 1996, for
a review). Most families live within 30 miles of their
institutionalized relative and within a median travel-
ing distance of 10 miles (Greene & Monahan, 1982;
Hook, Sobal, & Oak, 1982; Montgomery, 1982).
Findings regarding visitation frequency have been re-
markably consistent, with at least weekly visitation
for the majority (approximately 60%) of residents
(Bitzan & Kruzich, 1990; Hook et al., 1982; Na-
tional Center for Health Statistics, 1979). Nonethe-
less, large numbers of nursing home residents receive
less frequent visits. In one study, 16% received no
visits by family and 40% received no visits from
friends (Bitzan & Kruzich, 1990).
A handful of factors have been examined in rela-
tion to the disparity in contact levels of nursing home
residents. Contact is greater for residents whose care
is funded by family resources (Geib, 1980). Contact
also increases with proximity. As geographic dis-
tance of the nursing home to a family caregiver
decreases, visitation increases (Bitzan & Kruzich,
1990). Those with children living within an hour’s
drive receive more visits (Hook et al., 1982). Another
study found that for each doubling of distance be-
tween the nearest direct relative and the nursing
home, visitation declined by about half (Greene &
Monahan, 1982). Kinship status also is important:
more visits are received by persons who are married
(Hook et al., 1982; National Center for Health Sta-
tistics, 1979). Residents with greater impairments in
activities of daily living (ADLs) receive more visits
Key Words: Family involvement, Long-term care,
Dementia, Caregiving
Understanding family caregiving to frail elders is a
large and growing area within gerontology; however,
the attention paid to caregivers has generally ended
at the nursing home door. Consequently, little is
known about family caregiving to the 1.5 million
persons aged 65 and older currently residing in nurs-
ing homes in the United States (U.S. Department of
Health and Human Services, 1999). This gap in un-
This research was supported by Grant R01 AG8211 from the National
Institute on Aging. We acknowledge the cooperation of the facilities, resi-
dents, and families participating in the Maryland Long-Term Care Project.
Address correspondence to Cynthia L. Port, PhD, Department of Epi-
demiology and Preventive Medicine, University of Maryland, Baltimore,
660 West Redwood Street, Suite 200, Baltimore, MD 21201. E-mail:
cport@epi.umaryland.edu
1Department of Epidemiology and Preventive Medicine, University of
Maryland, Baltimore.
2School of Social Work, University of North Carolina at Chapel Hill.
Vol. 41, No. 5, 2001
589

(Greene & Monahan, 1982). No impact on contact
level has been found for visitor gender, education, or
retirement status, nor for the age of the resident
(Hook et al., 1982). The relationship of the duration
of stay in the nursing home to contact is unclear.
Some studies have found little change over time
(Greene & Monahan, 1982; Spasoff et al., 1978),
others have found decreased contact with increased
length of stay (Hook et al., 1982), and others have
arrived at mixed results depending on whether the
visitors are family members (decreased visits) or
friends (no change; Bitzan & Kruzich, 1990).
These findings begin to describe which residents
receive greater contact from family and friends, but
at least three potentially important factors have yet
to be examined: dementia status, race, and preadmis-
sion visitation. Little is known about the relationship
between contact level and the cognitive health of the
resident. Family and friends may mobilize around
demented elders, feeling a responsibility to more
closely monitor care and provide a connection to the
resident’s past. Alternatively, visiting someone with
dementia may be less personally rewarding, or more
emotionally distressing, leading to reduced contact.
Cognitively impaired residents can lose the ability to
participate in certain forms of contact, such as phone
calls and written communication.
Race/ethnicity has been shown to be related to
nursing home admission rates, with lower use of in-
stitutional care by older African Americans, Hispan-
ics, and Asian Americans as compared to Caucasians
(Damron-Rodriguez, Wallace, & Kington, 1994;
Falcone & Broyles, 1994). The larger social support
networks available to elderly Blacks (Gibson, 1982;
Gibson & Jackson, 1987; Thornton, White-Means,
& Choi, 1993) may be providing care longer and
thus delaying or avoiding institutionalization. Other
possible explanations for the reduced use of formal
long-term care services among Blacks may include
socioeconomic disadvantage, decreased access to long-
term care, and racial discrimination. Considering that
nursing home admission rates differ by race, it may
be useful to examine nursing home contact levels by
race as well.
Preadmission contact is another largely over-
looked but potentially important component of post-
admission contact. Family and friends who are
highly involved prior to placement may be more
likely to remain so after placement. An early study
suggests that telephone contact prior to placement is
weakly but positively related to later nursing home
visits (York & Caslyn, 1977).
work of family and friends. In one study, more than
three quarters of nursing home visits were by groups
of persons (Hook et al., 1982). A fuller understand-
ing of family involvement should encompass how
many persons maintain contact with a nursing home
resident via in-person visits or telephone calls.
The existing literature also has limited generaliz-
ability to today’s nursing home populations. Most
studies relied on small and nonrepresentative sam-
ples from rural areas, often selected from a limited
number and type of nursing homes. These studies ei-
ther include no information about participants’ race
or indicate that the sample was essentially all-White.
Residents were sometimes excluded on the basis of
illness and impaired cognition. Moreover, most of
the studies were published more than 15 years ago.
Given the increasing age and acuity of today’s nurs-
ing home residents, both the residents and their so-
cial networks may be different now.
The current study addressed these limitations in
several ways. Reports of in-person visits and tele-
phone calls were combined to create a more compre-
hensive measure of contact. Visits and phone calls
from spouses, siblings, children, other relatives, and
friends/neighbors were considered separately to ob-
tain a more complete picture of the contact patterns
of the resident’s entire social network. The study uti-
lized a large and representative sample. In addition
to reexamining several of the factors shown in previ-
ous studies to affect visitation rates, the current study
also examined preadmission contact levels and resi-
dent dementia status and race. Finally, we conducted
an analysis examining the relative contributions of
these factors to postadmission contact levels while
controlling for preadmission contact levels.
Based on previous research, we hypothesized the
following. Higher levels of postadmission contact
would be positively related to non-Medicaid pay-
ment status, closer kinship status, a greater number
of family and friends in geographic proximity to the
nursing home, and a higher number of ADL depen-
dencies. We expected that residents with dementia
would have less contact with family and friends. We
also hypothesized that Black nursing home residents
would receive higher contact levels than would
White residents. Finally, we expected contact levels
prior to admission to be positively correlated with
postadmission contact levels.
Methods
Participants and Procedure
Methodological limitations are evident in the ex-
isting literature on family involvement in the nursing
home. Nursing home contact is generally measured
as the number of in-person visits received by a resi-
dent. However, more distant relatives may utilize
other forms of communication such as telephone
calls. Contact has also been traditionally measured
as the visitation of a single primary significant other.
This approach underestimates contact levels for resi-
dents whose social support is shared across a net-
Subjects were interviewed as part of the Maryland
Long-Term Care Project (LTCP), a program of re-
search dedicated to studying quality of life among
persons residing in long-term care settings (Maga-
ziner et al., 2000). Study participants were the signif-
icant others (as identified by staff) of nursing home
residents enrolled in the LTCP. Nursing homes were
identified and residents were enrolled as follows. All
221 licensed long-term care facilities in Maryland
590
The Gerontologist

were categorized by location and number of beds,
and a stratified random sample of 64 homes was se-
lected. Four (6%) homes did not participate and one
had no eligible residents. The 59 homes providing
data represented a 27% sample of all Maryland
nursing facilities and included rural, urban, and sub-
urban homes as well as a range of nursing home
sizes. All residents newly admitted from September
1992 through March 1995 aged 65 years and older
who had not resided in a nursing home in the previ-
ous year (for more than 8 days) were eligible. Sev-
enty percent of the 3,851 eligible residents were en-
rolled, for a total of 2,285. Most residents were
female (70.9%), the average age was 81.5 years,
80.5% (1,837) were White, 15.7% were Black,
0.9% were Hispanic, and 2.8% were described as
“other.” These demographics are comparable to
national figures (Magaziner et al., 2000; Strahan,
1997). Efforts were made to interview one significant
other per resident over the telephone or in person;
2,112 were interviewed.
Only interviews with family members and friends
(excluding guardians and paid helpers) 18 years or
older that had taken place at least 2 weeks (14 days)
after the resident’s admission to the nursing home
(for at least a 15-day stay) were included in the cur-
rent sample (N ꢀ 1,441). Guardians and paid help-
ers (n ꢀ 73) were excluded out of concern that they
would be unfamiliar with the contact patterns of the
resident’s family and friends. Of the significant oth-
ers included in the present analyses, most were fe-
male (n ꢀ 979; 67.9%) and White (n ꢀ 1,211;
84%). They had a mean age of 57.4 years (range 25–
98; SD ꢀ 12.6), 13.6 years of education (range 0–20;
SD ꢀ 3.1), and were interviewed an average of 34.3
days following the resident’s admission (range 15–65;
SD ꢀ 13.0). The significant others had known the
patient an average of 51 years (range 2–90; SD ꢀ
14.2). The largest group was daughters (36.3%),
followed by sons (20.9%), spouses (12.8%), nieces
and daughters-in-law (5.5% and 5.3%, respectively),
granddaughters and sisters (3.3% each), and friends
and neighbors (3.2%). The remainder was made up
of other family relationships. Resident characteristics
for the final sample were very similar to the full sam-
ple: average age 81.7 years (range 65–99; SD ꢀ 7.6),
72.2% female, and 81% (n ꢀ 1,167) White. Of non-
Whites in the sample, 15% (n ꢀ 216) were Black,
1.1% (n ꢀ 16) were Hispanic, and 2.9% (n ꢀ 42)
were “other.” For this reason, only White versus
Black comparisons were conducted in the regression
analysis, excluding the Hispanic and “other” catego-
ries.
“How many (brothers/sisters) live within a one-hour
drive of the nursing home?” Significant others were
also asked how many miles the nursing home was
from the resident’s nearest family member or friend
(nearest loved one), how many miles the nursing
home was from the home in which the resident lived
prior to admission (geographic relocation), and how
many relatives lived within a one-hour drive of the
resident prior to admission, including those living
with the resident (preadmission network proximity).
Contact.—Contact was defined as the number of
days in a defined 2-week period that a resident was
visited or telephoned by specific categories of per-
sons. Preadmission contact was the total number of
visits and calls by four categories of persons who did
not live with the resident: siblings, children, other
relatives, and friends/neighbors. Persons living in the
same household with the resident were automatically
assigned daily contact for visits. For postadmission
contact, information was obtained regarding five cat-
egories of persons: spouse, siblings, children, other
relatives, and friends/neighbors. Preadmission con-
tact covered the 2 weeks before the resident was ad-
mitted to the nursing home. Postadmission contact
referred to the 2 weeks preceding the interview with
the significant other.
Medicaid Status.—Residents whose stay in the
nursing home was funded in some part by Medicaid
received a 1 for this variable, while all others re-
ceived a 0. Data were obtained from the Minimum
Data Set (MDS; Morris et al., 1990) upon admission.
Number of ADL Dependencies.—Current physical
functioning was rated by nursing home staff with a
modified version of the Katz Activities of Daily Liv-
ing Scale (Katz, Ford, & Moskowitz, 1963). The
scale was derived from 14 items in the Psychogeriat-
ric Dependency Rating Scale (PGDRS; Wilkinson
& Graham-White, 1980) on six domains: bathing,
dressing, toileting, transferring, feeding, and conti-
nence. Each Katz domain was scored dichotomously
as any dependency versus fully independent. A sum-
mary Katz score was created by adding the number
of dependent domains; higher scores are indicative of
more dependence.
Kinship Status.—Residents whose significant other
was a first-degree family member (sibling, child, par-
ent, or spouse) were assigned a 1 for Kinship status.
Residents whose significant other was not a first-
degree relative were assigned a 0.
Variables
Network Proximity.—A count of the number of
persons living within a one-hour drive of the nursing
home was obtained for five categories of persons:
spouse, siblings, children, other relatives, and friends/
neighbors (postadmission network proximity). For
example, for each category, respondents were asked,
Dementia Status.—Dementia status was deter-
mined in accordance with the Diagnostic and Statis-
tical Manual of Mental Disorders, 3rd edition, re-
vised (American Psychiatric Association, 1987) by an
expert panel of geriatric psychiatrists, neurologists,
and a geriatrician using information collected from
Vol. 41, No. 5, 2001
591

interviews with residents, nursing staff, significant
others, and medical records. A detailed description
of the dementia ascertainment methodology may be
found elsewhere (Magaziner et al., 1996). For the
present analyses, residents with dementia were as-
signed a 1, and those determined not to have demen-
tia (or whose dementia status could not be deter-
mined conclusively) were assigned a 0.
mission. Approximately half (53.9%, n ꢀ 776) of
the residents had dementia.
Relationship to Postadmission Contact
Hypotheses regarding the relationships between
contact and the primary study variables were exam-
ined via correlations for continuous variables and t
tests for categorical variables. Table 2 shows the cor-
relations. Contrary to prediction, the number of
ADL dependencies was not significantly correlated
with postadmission contact. As predicted, preadmis-
sion contact was significantly positively correlated
with postadmission contact (r ꢀ .35, p ꢁ .001), indi-
cating that residents whose social networks offered
high levels of contact prior to admission received
more visits and phone calls than residents whose so-
cial networks provided lower contact prior to admis-
sion. Postadmission network proximity was also pos-
itively correlated with post-admission contact (r ꢀ
.23, p ꢁ .001); residents with a larger number of rel-
atives and friends living within a one-hour drive of
the nursing home had higher levels of contact.
Also as predicted, t tests indicated that residents
whose stay was paid at least in part by Medicaid re-
ceived less contact (M ꢀ 16.8, SD ꢀ 13.4) than those
who were not covered by Medicaid (M ꢀ 20.7, SD ꢀ
14.8), t(1350) ꢀ 4.14, p ꢁ .001; residents whose pri-
mary visitor was a first-degree relative had higher
contact levels (M ꢀ 21.2, SD ꢀ 15.2) than those
whose primary visitor was not a first-degree relative
(M ꢀ 15.9, SD ꢀ 12.1), t(1430) ꢀ –6.01, p ꢁ .001;
and residents with dementia had less contact (M ꢀ
17.3, SD ꢀ 12.5) than those not meeting criteria for
dementia (M ꢀ 22.9, SD ꢀ 16.4), t(1430) ꢀ –7.30,
p ꢁ .001). Finally, and contrary to expectations,
White residents had more contact (M ꢀ 20.5, SD ꢀ
14.3) than did Black residents (M ꢀ 16.2, SD ꢀ
14.9), t(1372) ꢀ 4.04, p ꢁ .001.
Results
Descriptive Data for Primary Study Variables
Table 1 provides descriptive information about
residents and significant others. On average, nursing
home residents lived less than 15 miles from a signif-
icant other. Mean relocation following placement
was less than 20 miles. The mean number of ADL
skills in which the resident was dependent was close
to four. The number of family members and friends
living within an hour’s drive did not change follow-
ing the resident’s placement; the median number
both pre- and postadmission was 7. The mean num-
ber of contacts by family members and friends who
visited, called, or lived with the resident in the
2-week period prior to placement (37.8) compared
to the mean number of contacts during the 2-week
period preceding the significant other interview
(19.9) decreased by approximately half. The median
number of contacts decreased from 33 prior to place-
ment to 17 following placement. The mean number
of visits dropped by 45% following admission; the
mean number of phone calls dropped by 56%.
The primary significant other was a first-degree
relative for 1,087 (75.4%) of the residents. A total of
315 (23.2%) of the residents utilized Medicaid at ad-
Table 1. Descriptive Statistics for Primary Study Variables
Dementia Status and Race
Variable
M
Median SD Range
n
In an effort to better understand differences in
contact levels by dementia status and race, we exam-
ined relationships between these variables and the re-
maining primary study variables.
Nearest loved one (miles) 13.51
Geographic relocation
6.0
52.1 0–900 1,424
68.8 0–900 1,423
(miles)
18.4
3.9
6.0
4.0
ADL dependencies
Preadmission network
proximity (# family
and friends within
one-hour drive)
Postadmission network
proximity (# family
and friends within
one-hour drive)
1.9 0–6
1,238
Dementia Status.—Residents with dementia were
different from those without dementia in several
ways. Chi-square tests showed that demented resi-
dents were significantly more likely to have some
part of their care paid for by Medicaid (p ꢁ .001;
17.4% for those without dementia and 28% for
those with dementia). A series of t tests showed that
demented residents were more dependent in ADLs
(M ꢀ 4.3, SD ꢀ 1.7) than nondemented residents (M ꢀ
3.3, SD ꢀ 2.0) t(1236) ꢀ 9.29, p ꢁ .001; had fewer
relatives and friends living within a one-hour drive of
their home (M ꢀ 10.3, SD ꢀ 14.4) than nonde-
mented residents (M ꢀ 12.0, SD ꢀ 17.6), t(1438) ꢀ
–2.00, p ꢁ .05; and had fewer relatives and friends
living within a one-hour drive of the nursing home
(M ꢀ 10.2, SD ꢀ 13.7) than nondemented residents
11.0
11.0
7.0
7.0
15.9 0–190 1,440
15.4 0–190 1,438
Preadmission contact
(per 2 weeks)
37.8 33.0
28.6 25.0
23.6 0–149 1,435
20.9 0–140 1,441
Preadmission Visits
Preadmission phone
calls
9.2
5.0
11.0 0–56 1,435
Postadmission contact
(per 2 weeks)
19.9 17.0
15.8 14.0
14.7 0–112 1,432
10.2 0–66 1,440
Postadmission Visits
Postadmission phone
calls
4.1
0
8.8 0–59 1,432
Note: ADL ꢀ activity of daily living.
592
The Gerontologist

Table 2. Pearson Correlations Among Primary Continuous Study Variables
Variable
1
2
3
4
5
6
7
1. Nearest loved one
2. Geographic relocation
3. ADL dependencies
—
.04
—
ꢃ.01
ꢃ.03
—
ꢃ.04
ꢃ.02
.01
ꢃ.05
.04
.01
ꢃ.01
.01
.08**
ꢃ.09**
ꢃ.01
ꢃ.04
4. Preadmission network proximity
5. Postadmission network proximity
6. Preadmission contact
—
.90***
—
.30***
.24***
—
.21***
.23***
.35***
—
7. Postadmission contact
Note: n ꢀ 1,222 to 1,438; ADL ꢀ activity of daily living.
**p ꢁ .01; ***p ꢁ .001.
(M ꢀ 11.9, SD ꢀ 17.1), t(1436) ꢀ –2.06, p ꢁ .05.
Residents with and without dementia did not differ
statistically in terms of kinship status or amount of
contact prior to nursing home admission.
ling for preadmission contact, variables shown in
previous studies to be related to contact also contrib-
uted significantly to the prediction of postadmission
contact. Both Medicaid status and ADL dependen-
cies were significant contributors, with Medicaid sta-
tus and greater ADL dependency predicting reduced
contact. Kinship status and postadmission network
proximity were also significant predictors of postad-
mission contact. Specifically, residents whose signifi-
cant other was a first-degree relative and residents
with a larger social network living within an hour’s
drive of the nursing home experienced more postad-
mission contact. In the third and final step, two of
the new variables examined in this study, dementia
status and race, also contributed significantly to the
prediction of postadmission contact. Results of the re-
gression model were virtually identical when the
MDS cognitive score was utilized in place of the di-
chotomous dementia status variable: adjusted R2 ꢀ
.13 for Step 1; ꢂ adjusted R2 ꢀ .06 for Step 2 (p ꢁ
.001); ꢂ adjusted R2 ꢀ .03 for Step 3 (p ꢁ .001).
Both being demented and being Black were associ-
ated with less postadmission contact. Moreover, af-
ter including dementia status and race in the model,
neither Medicaid status nor greater ADL impairment
Race.—Black and White residents also differed
from each other in important ways. First, 51.1% (n ꢀ
596) of White residents and 68.5% (n ꢀ 148) of
Black residents met criteria for dementia (p ꢁ .001).
Second, whereas 77.8% (n ꢀ 908) of significant oth-
ers of White residents were a direct relative, this was
the case for 60.2% (n ꢀ 130) of Black residents (p ꢁ
.001). Third, whereas 18% (n ꢀ 197) of White resi-
dents utilized Medicaid to finance some part of their
nursing home care, 49.3% (n ꢀ 103) of Black resi-
dents did so (p ꢁ .001). A t test also demonstrated
that Blacks had more ADL dependencies (M ꢀ 4.3,
SD ꢀ 1.8) than Whites (M ꢀ 3.8, SD ꢀ 1.9), t(1191) ꢀ
3.46, p ꢁ .01. However, Blacks had a greater num-
ber of persons living within an hour of their home
before admission (M ꢀ 14.2, SD ꢀ 25.9) than did
Whites (M ꢀ 10.5, SD ꢀ 13.6), t(1380) ꢀ 3.06, p ꢁ
.01; a greater number of persons living within an
hour of the nursing home (M ꢀ 14.1, SD ꢀ 24.7)
than did Whites (M ꢀ 10.4, SD ꢀ 13.2), t(1378) ꢀ
3.20, p ꢁ .01; and a greater amount of contact prior
to nursing home admission (M ꢀ 42.6, SD ꢀ 28.1)
than did Whites (M ꢀ 36.7, SD ꢀ 22.5), t(1375) ꢀ
3.35, p ꢁ .01).
Table 3. Regression Models Assessing the Effect of the Primary
Study Variables on Postadmission Contact
Variable
Step 1
Preadmission contact
Step 2
Preadmission contact
b
SE b
Beta
Predictors of Postadmission Contact Controlling for
Preadmission Contact
Regression models were used to explore the rela-
tive contributions of resident and significant other
characteristics to postadmission contact levels. In the
first model, preadmission contact was regressed onto
postadmission contact. In the second model, vari-
ables previously found to be related to postadmission
contact were entered: Medicaid status, kinship sta-
tus, postadmission network proximity, and the num-
ber of the resident’s ADL dependencies. In the final
model, two new variables examined in this study
were entered, namely, dementia status and race
(White vs Black). Results of the regression modeling
are shown in Table 3.
0.21
0.19
ꢃ3.3
4.5
0.13
.02
.02
.35***
.32***
Medicaid status (Medicaid)
Kinship status (first degree)
Postadmission network proximity
ADL dependencies
.94 ꢃ.10**
.93
.02
.13***
.15***
ꢃ0.68
.21 ꢃ.09**
Step 3
Preadmission contact
0.19
ꢃ1.9
3.9
.02
.97 ꢃ.06
.93
.02
.21 ꢃ.04
.84 ꢃ.15***
1.1
.32***
Medicaid status (Medicaid)
Kinship status (first degree)
Postadmission network proximity
ADL dependencies
Dementia status (Demented)
Race (Black)
.12***
.14***
0.13
ꢃ0.33
ꢃ4.3
ꢃ3.8
ꢃ.10**
In the first model, preadmission contact accounted
for a substantial proportion of variance in postad-
mission contact. In the second model, after control-
Note: Adjusted R2 ꢀ .12 for Step 1; ꢂ Adjusted R2 ꢀ .05 for
Step 2 (p ꢁ .001); ꢂ Adjusted R2 ꢀ .03 for Step 3 (p ꢁ .001).
**p ꢁ .01; ***p ꢁ .001.
Vol. 41, No. 5, 2001
593

was significant, although Medicaid status approached
significance. In terms of their relative contributions
to the prediction of postadmission contact in the
final regression equation, postadmission network
proximity and dementia status followed preadmis-
sion contact in importance, followed by kinship sta-
tus and race. The total amount of variance in postad-
mission contact accounted for by the final model was
modest at 21%.
frequency of contact prior to admission and their
likelihood of having a first-degree relative as a pri-
mary significant other did not differ from nonde-
mented residents, however. These results suggest that
residents with dementia begin with a smaller avail-
able social network than that of the nondemented
and also suffer greater decreases in contact, perhaps
related to financial barriers.
Interpersonal issues, though not examined here,
may also lead to decreased contact for residents with
dementia. The study by York and Caslyn (1977)
found enjoyment of visits decreased as resident self-
care ability, cognitive functioning, and personal
appearance declined. Moreover, families reported
greater difficulty coping with mental deterioration
(37%) than with physical disabilities (15%). Resi-
dent cognitive impairment was also shown to be re-
lated to poorer nursing care and reduced staff aware-
ness of medical problems (Fleishman, Rosin, Tomer,
& Schwartz, 1987). This disturbing finding may re-
late to the resident’s reduced ability to direct his or
her care, but may also reflect a lack of involved fam-
ily members to advocate on the resident’s behalf. Re-
search is needed on the reactions of family and
friends to the illness and impairment of nursing
home residents and how this relates to contact levels.
Another variable examined in the current study
for the first time is race. Black residents were hypoth-
esized to have greater contact from family and
friends than Whites, but this was not the case. Black
residents did have greater contact than Whites prior
to admission and, as seen in other studies, Black fam-
ily networks were larger in size and more diverse in
terms of kinship than those of Whites (Burton et al.,
1995; Gibson, 1982; Gibson & Jackson, 1987;
Thornton et al., 1993). The finding of reduced con-
tact following admission therefore suggests that bar-
riers to the continuation of contact may be greater
for Blacks than for Whites. One barrier may be
lower family resources, as Black residents were sig-
nificantly more likely to utilize Medicaid than were
White residents. Lower socioeconomic status may
translate into greater work obligations that diminish
opportunities to visit during facility visiting hours,
difficulty paying for transportation and telephone
costs, and many other practical barriers. Black resi-
dents were also more dependent in ADLs and more
likely to have dementia than White residents, both
of which have been documented previously among
nursing home residents (Engle & Graney, 1995).
This suggests that the residents may have been cared
for in the community for a longer time and only
brought to a nursing home when the family’s ability
to continue care had diminished. Finally, Black resi-
dents were less likely to have a first-degree relative as
a primary significant other. Burton and colleagues
(1995), noting racial differences in support systems
for elders, point out that little is known about the
stability of more diverse family systems over time.
Variables not examined in the current study, but
which might also pertain to Black elders and their
families, include the overall health of the social net-
Discussion
This study examined factors related to the level of
contact between nursing home residents and their
family and friends. In addition to reexamining vari-
ables previously shown to be related to contact, three
new variables were examined, and the relative con-
tributions of the study variables in predicting postad-
mission contact while controlling for preadmission
contact were analyzed. The study is unique with re-
gard to the size and representativeness of the sample,
the expanded definition of contact, and the inclusion
of the resident’s larger social network.
With one exception, our results confirmed previ-
ous findings. Residents receiving Medicaid assistance
had less contact with family and friends. Greater
monetary resources may facilitate contact in several
ways; for example, family members of higher socio-
economic status may experience fewer practical bar-
riers to visitation and phone calls. Residents whose
primary significant other was a first-degree relative
also had higher contact levels. Stronger emotional
bonds and/or greater sense of familial duty likely ex-
ist between closely related family members as op-
posed to those more distantly related, or those re-
lated only by marriage.
The number of family and friends living in close
proximity to the nursing home was also related to
more contact. Although network proximity is a mea-
sure of access to family and friends, it is also an esti-
mate of the social network size of the resident. Most
likely both the nearness and sheer number of family
and friends relate positively to contact levels. Studies
that specifically measure the impact of social net-
work size on contact levels would be informative in
this regard.
One finding from previous research was not borne
out in our study; ADL dependencies were not corre-
lated with postadmission contact. In the regression
analysis, ADL dependencies lost significance when
dementia status and race were considered. This find-
ing is perhaps not surprising considering that ADL
dependencies are higher among demented elders than
among the nondemented, and being nondemented
was a strong predictor of postadmission contact.
Dementia status has not previously been studied in
relation to family contact patterns, despite its obvi-
ous relevance to nursing home populations. This
study found that residents with dementia receive less
attention from family and friends. Demented resi-
dents were also more likely to utilize Medicaid and
to have fewer family and friends living nearby. Their
594
The Gerontologist

work, additional filial duties, and cultural differences
such as level of comfort in visiting a nursing home.
Research focused on barriers to visitation that differ-
entially affect Black families is needed.
included only newly admitted residents, we must be
cautious in interpreting the results. Contact patterns
may change to a greater degree at times of transition,
such as just prior to and just after nursing home ad-
mission. On the other hand, the use of new admis-
sions was valuable in that it facilitated more accurate
recollections of preadmission contact levels. Given
the social expectation that contact levels be high, it is
possible that the reports from significant others used
here were inflated. Unfortunately, no clearly superior
alternative exists for measuring contact levels. Staff
reports also entail biases and inaccuracies, and direct
observation to count all of the visits and phone calls
received by a resident would be logistically very diffi-
cult. Because there is no reason to suppose that some
population groups would be more likely than others
to inflate contact rates, biases that exist would tend
toward the null.
Study results have implications for the design of
interventions to improve family involvement in the
nursing home. Preadmission contact would clearly be
difficult to alter. Likewise, neither increasing the
number of persons within a resident’s support net-
work nor increasing the likelihood that a primary
significant other is a first-degree relative are feasible.
While one can recommend that elders be placed in
nursing homes that are near to their support net-
works, it appears from this study that this goal is al-
ready being accomplished. An approach with more
potential for improvement may be to target residents
with dementia. Loved ones may need more informa-
tion about dementia in order to increase their com-
fort level while visiting a cognitively impaired resi-
dent. Such an intervention could also emphasize the
critical need for family and friends to advocate on
behalf of a person with dementia as his or her own
ability to do so lessens. Socioeconomic factors may
also be amenable to intervention. For example,
transportation barriers may be ameliorated through
volunteer ride systems; less flexible work schedules
may be accommodated through longer visiting hours,
and addressing the psychiatric and health needs of
significant others may improve their ability to be in-
volved in the nursing home. There is currently insuf-
ficient information regarding the salience of these
barriers or what steps would be effective in reducing
their negative effect on contact levels.
Perhaps the most striking finding of the present
study is that, although families tend to live quite
close to their loved one in a nursing home, contact
via in-person visits and telephone calls dropped by
approximately half within the first few weeks and
months of placement. Although this does not indi-
cate that families are no longer involved in care (resi-
dents received visits or calls from 20 persons per
2-week period on average), the decline is striking.
Some drop in contact would be expected as persons
who lived in the same household with the resident
prior to admission (and were automatically coded as
having a daily visitation contact rate) must now go
to the home to visit. However, drops in contact are
similar for visits and phone calls. Postadmission con-
tact was not related to the distance the resident
moved when admitted to a nursing home; however,
residents generally moved very short distances from
home. To our knowledge, this study is the first to
closely examine change in the amount of family in-
volvement from pre- to postadmission.
This study also provides one of the most complete
explorations to date of how previous family habits
and demographic, health, and socioeconomic factors
relate to contact after admission. Preadmission con-
tact accounted for the largest portion of variance in
explaining postadmission contact, with higher pread-
mission contact positively related to postadmission
contact. The longstanding contact patterns of an in-
dividual’s social network are clearly important to
keep in mind when evaluating patterns of contact af-
ter admission. Other variables of importance are the
proximity of family and friends, kinship status of the
primary significant other, dementia status, and race.
Nevertheless, the overall variance accounted for by
the regression equation is modest (approximately
21%), indicating that much more has yet to be
learned about families and nursing home contact.
Other variables to consider include (a) the health of
the resident’s social network, (b) the extent of addi-
tional time constraints such as employment and
other caregiving obligations, (c) transportation and
other practical barriers, (d) the quality of the per-
sonal relationship between the resident and his or her
support system, (e) the quality of the personal rela-
tionships between members of the resident’s support
system, and (f) characteristics that make the nursing
home facility a welcoming setting for maintaining in-
volvement.
References
American Psychiatric Association. (1987). Diagnostic and statistical man-
ual of mental disorders (3rd ed., rev.). Washington, DC: Author.
Bitzan, J. E., & Kruzich, J. M. (1990). Interpersonal relationships of nurs-
ing home residents. The Gerontologist, 30, 385–390.
Burton, L., Kasper, J., Shore, A., Cagney, K., LaVeist, T., Cubbin, C., &
German, P. (1995). The structure of informal care: Are there differ-
ences by race? The Gerontologist, 35, 744–752.
Damron-Rodriguez, J., Wallace, S. P., & Kington, R. (1994). Service utili-
zation and minority elderly: Appropriateness, accessibility and
acceptability. Gerontology and Geriatrics Education, 15, 45–64.
Engle, V. F., & Graney, M. J. (1995). Black and white female nursing
home residents: Does health status differ? Journal of Gerontology:
Medical Sciences, 50A, M190–M195.
Falcone, D., & Broyles, R. (1994). Access to long-term care: Race as a
barrier. Journal of Health, Political Policy, and Law, 19, 583–595.
Fleishman, R., Rosin, A., Tomer, A., & Schwartz, R. (1987). Cognitive
Although addressing many shortcomings of the
existing literature, the current study also has limita-
tions, most importantly the short length of nursing
home stay, the use of newly admitted residents only,
and the use of reports of contact frequency from sig-
nificant others. Length of stay has not been conclu-
sively found to be associated with a decrease in con-
tact, but longitudinal studies are needed to assess
change over longer time periods. Because our study
Vol. 41, No. 5, 2001
595

impairment and the quality of care in long-term care institutions.
Comprehensive Gerontology, B, 1, 18–23.
Geib, H. F. (1980). Visitation patterns and economic support of nursing
home patients. The Gerontologist, 20, 108–111.
Gibson, R. C. (1982). Blacks at middle late life: Resources and coping. An-
nals of the American Academy of Political and Social Science, 464,
79–90.
Gibson, R. C., & Jackson, J. S. (1987). Health, physical functioning, and
informal supports of the black elderly. Milbank Quarterly, 65
(Suppl.), 421–454.
Greene, V. L., & Monahan, D. J. (1982). The impact of visitation on pa-
tient well-being in nursing homes. The Gerontologist, 22, 418–423.
Hook, W. F., Sobal, J., & Oak, J. C. (1982). Frequency of visitation in
nursing homes: Patterns of contact across the boundaries of total insti-
tutions. The Gerontologist, 22, 424–428.
Katz, D., Ford, A. B., & Moskowitz, R. W. (1963). Studies of illness in the
aged. The index of ADL: a standardized measure of biological and
psychological function. Journal of the American Medical Association,
185, 914–919.
Magaziner, J., German, P., Zimmerman, S. I., Hebel, J. R., Burton, L.,
Gruber-Baldini, A. L., May, C., & Kittner, S. (2000). The prevalence
of dementia in a statewide sample of new nursing home admissions
age 65 and older: Diagnosis by expert panel. The Gerontologist, 40,
663–672.
Magaziner, J., Zimmerman, S. I., German, P., Kuhn, K., May, C., Hooper,
F., Cox, D., Hebel, J. R., Kittner, S., Burton, L., Fishman, P., Kaup, P.,
Rosario, J., & Cody, M. (1996). Ascertaining dementia by expert
panel in epidemiological studies of nursing home residents. Annals of
Epidemiology, 6, 431–437.
the national resident assessment instrument for nursing homes. The
Gerontologist, 30, 293–307.
National Center for Health Statistics. (1979). The National Nursing
Home Survey: 1977 summary for the United States. Vital and Health
Statistics, Series 13, No. 43, (DHEW Publication No. [PHS] 79-1794).
Washington, DC: U.S. Government Printing Office.
Naleppa, M. (1996). Families and the institutionalized elderly: A review.
Journal of Gerontological Social Work, 27(1/2), 87–111.
Spasoff, R. A., Kraus, A. S., Beattie, E. J., Holden, D., Lawson, J. S.,
Rodenburg, M., & Woodcock, G. W. (1978). A longitudinal study of
elderly in long stay institutions. 1. Early response to institutional care;
2. The situation one year after admission. The Gerontologist, 18, 281–
292.
Strahan, G. W. (1997). An overview of nursing homes and their current
residents: Data from the 1995 National Nursing Home Survey. Ad-
vance Data, 23(280), 1–12.
Thornton, N. C., White-Means, S. I., & Choi, H. K. (1993). Sociodemo-
graphic correlates of the size and composition of informal caregiver
networks among frail ethnic elderly. Journal of Comparative Family
Studies, 24, 235–250.
U.S. Department of Health and Human Services. (1999). Health, United
States. 1999 Health and aging chartbook (DHHS Publication No.
[PHS] 99-1232-1). Washington, DC: U.S. Government Printing Office.
Wilkinson, I. M., & Graham-White, J. (1980). Psychogeriatric depen-
dency rating scales (PGDRS): A method of assessment for use by
nurses. British Journal of Psychiatry, 137, 558–565.
York, J. L., & Caslyn, R. J. (1977). Family involvement in nursing homes.
The Gerontologist, 17, 500–505.
Montgomery, R. (1982). Impact of institutional care policies on family in-
tegration. The Gerontologist, 22, 54–58.
Morris, J. N., Hawes, C., Fries, B. E., Phillips, C. D., Mor, V., Katz, S.,
Murphy, K., Drugovich, M. L., & Friedlob, A. S. (1990). Designing
Received December 20, 2000
Accepted June 4, 2001
Decision Editor: Laurence G. Branch, PhD
596
The Gerontologist