Journal of Gerontology: MEDICAL SCIENCES
2005, Vol. 60A, No. 11, 1469–1474
Copyright 2005 by The Gerontological Society of America
Strategies, Time, and Costs Associated With the
Recruitment and Enrollment of Nursing Home Residents
for a Micronutrient Supplementation Clinical Trial
Paula M. Gismondi,1 Davidson H. Hamer,1,2,3 Lynette S. Leka,1 Gerard Dallal,1
Maria A. Fiatarone Singh,4 and Simin N. Meydani1
1
Nutritional Immunology Laboratory Jean Mayer/USDA Human Nutrition Research Center on Aging (HNRCA),
Tufts University, Boston, Massachusetts. 2Division of Geographic Medicine and Infectious Diseases,
Tufts-New England Medical Center, Boston, Massachusetts. 3Center for International Health and Development,
Boston University School of Public Health, Massachusetts. 4University of Sydney, Lidcombe, Australia.
Background. Concomitant with the substantial growth of the elderly population in the last decade, there has been
a steady rise in the number of nursing home residents aged 65 years and older. Well designed, rigorously conducted
clinical intervention trials provide an important source of data for evidence-based improvements in the medical care of
nursing home residents. The information available on strategies for the recruitment and screening of participants
for such studies in long-term care facilities, as well as the financial and time costs for carrying out these investigations,
is limited.
Methods. This report describes our experience in recruiting 617 nursing home residents for a multisite, double-blind,
randomized, placebo-controlled trial designed to determine the efficacy of a 1-year period of vitamin E supplementation in
preventing respiratory tract infections. Comparisons of the projected staffing costs and actual costs incurred are presented,
using a retrospective method for the determination of unit costs.
Results. Initially, 874 consents were obtained from 2815 potential participants, of which only 617 were enrolled. Each
successful enrollment required an average of 15 hours of staff time at a combined personnel and supply cost of $515 per
participant and a total study cost of $317,661. Several obstacles were encountered during the recruitment and enrollment
process: resistance on the part of family or primary care provider; transfer out of facility; and changes in the medical
condition of the patient, including death.
Discussion. The results of this report should prove useful to investigators developing budgets for nursing home–based
clinical trials by providing a more accurate determination of the personnel needed and the costs associated with
recruitment and enrollment of participants.
M
ORE than 40% of the U.S. population over the age of
65 years will spend some time in a nursing home (1).
In 1997, on an average day, there were approximately 1.5
million elderly persons residing in nursing homes (2).
Physicians and nurses challenged with providing care to
elderly persons often lack scientifically valid data on which
to base diagnostic and therapeutic decision making. To
address the relative paucity of scientific evidence for prevention and treatment strategies in nursing home residents,
there is a need for well designed, rigorously conducted intervention trials.
Although there exists useful advice regarding ethical and
practical issues associated with the recruitment of nursing
home residents into clinical trials (3–5), little information is
available for the required staff time and other costs
associated with it. These estimates are necessary for the
successful implementation of clinical trials in nursing homes.
We recently completed a randomized, placebo-controlled
clinical trial designed to determine the effect of vitamin E
supplementation on the incidence and duration of respiratory tract infections in elderly nursing home residents
(6). We estimated recruitment costs in the original proposal
based on the limited literature and our experience conduct-
ing nutritional research in healthy, community-dwelling
elderly persons. As the study progressed, it became apparent that the personnel and costs needed to complete this trial
had been largely underestimated. The purpose of this article
is to provide the research community with data for more
accurately projecting recruitment costs and personnel time
for clinical intervention trials in elderly nursing home
residents.
STUDY METHODS
From 1988 through 2000, 617 participants from 33 longterm care facilities in the Boston area were enrolled in a 1year randomized, double-blind, placebo-controlled study of
vitamin E for the prevention of respiratory tact infections in
elderly persons. For details of design, study procedures, and
resident eligibility see Meydani and colleagues (6).
Identification and Recruitment of Nursing Homes
Prior to submission of the grant proposal, 15 Bostonarea nursing homes were surveyed. We learned that most
nursing homes would be receptive to the study as it was:
1) relatively noninvasive; 2) associated with minimal risk
and possible benefit to the participants; 3) minimally
1469
1470
GISMONDI ET AL.
burdensome for and at no cost to nursing home staff; and 4)
potentially able to associate the home with an academic institution. However, due to high turnover of administrators and nursing personnel, only three facilities from the
original sample participated in the study after it was funded.
The marketing approach to recruitment of homes for the
study arose from a literature review (3–5) and our own
observations during initial site visits. Resources to identify
homes included assistance from an aftercare coordinator at
New England Medical Center, nursing home directories
published by the Massachusetts Department of Public
Health (7), and our personal and professional associates.
Large long-term care facilities geographically close to our
research center located in downtown Boston and smaller
homes located within the city’s ethnic neighborhoods were
chosen for our marketing campaign. A marketing packet
that included a cover letter, study brochure, abstract, and
reprints of pertinent publications by the principal investigator, was mailed to the administrators and directors of nursing
and medicine at each facility. Two weeks after the mailings,
a follow-up phone call was made to the nursing director to
determine interest and to schedule an informational meeting.
Successful meetings produced: 1) a letter of cooperation
signed by the homes’ administrator; 2) a letter co-signed by
the homes’ medical director and study physician informing
the primary care providers (PCPs) of the study requirements; and 3) a scheduled date for in-service training with
the homes’ nursing staff.
Participant Recruitment
In-service sessions with staff nurses were held at each site
to inform them of the study’s significance, possible impact
on health status of all elderly persons, study procedures, and
eligibility criteria. Additionally, an overview of the research
institution and qualifications of the research team, the
noninvasive nature of the study, and the limited amount of
work that would be imposed on the nursing home staff were
emphasized. Assurance was given that staff was not expected to obtain informed consent or to perform additional
documentation other than recording the daily supplement
administration. Multitest CMI skin test kits (Merieux Institute, Inc., Miami, FL), used to determine immune response
as part of the study objectives, and sample supplements
were displayed at each meeting. Ample time was allowed
to address any misperceptions or problematic past experiences that the nurses might have had with medical
research.
Following each in-service session, the research team and
nursing staff reviewed the resident list for potential
eligibility. The research team briefly reviewed each
candidate’s clinical chart for contact information to notify
families and the PCP that the patient was being considered
for the study. A spreadsheet of patient information was
created so that the research team could track the status of
each potential candidate.
Notification to the PCP
The first step in the recruitment process was to send
a letter, listing each patient candidate and co-signed by the
medical director and a study physician, to each PCP. The
letter included a description of the study and outlined its
impact on the current care of the patient should the decision be made to participate. The PCP was advised to contact
the nursing director if the patient was not a good candidate.
The recruitment team waited 2 weeks to see whether the
PCP objected before approaching patients or their proxies
for consent.
Determining Competency of Eligible Patients
Nursing home staff made a preliminary determination of
competency; however, further evaluation by the research
team of the patient’s orientation, memory, and judgment
was required. Any ambivalence regarding these cognitive
markers led the research team to obtain a consent co-signed
by the candidate’s guardian or closest family member.
Informed Consent
The research team visited all candidates who were considered competent and explained the study in general terms to
ascertain interest and willingness to participate. Those
interested were given the consent form to read. The research
team read the consent aloud to those who were sight impaired.
The resident was queried throughout the interview to ensure
that the study purpose and procedures were understood.
The nursing home staff was hesitant to serve as witness to
the consent process. Therefore, the presence of two research
team members was required, one to review the consent form
and the other to witness the process.
Letters co-signed by the nursing home administrator and
Directors of Nursing and Medicine were sent to the next of
kin or guardian of patients considered incompetent,
informing them that the home was participating in the
study, their relative was a potential study candidate, and
a member of the research team would be contacting them.
Enclosed with the letter were a study brochure, informed
consent form, and a postage-paid return envelope.
Follow-up phone calls were made 2 weeks after the
mailing. Repeat phone calls often had to be made if the
family members had not been available, wanted time to
speak to siblings about the study, or had misplaced or never
received the materials. In most cases more than one mailing
and several phone calls were required to ascertain the
family’s decision.
Over the phone, the research staff encouraged the family
to ask questions about the procedures and risks involved.
Those who agreed to enrollment were asked to sign the
consent, have their signatures witnessed, and mail back the
form in the return envelope. The consent forms included
specific signature lines that denoted the family members’
relationship to the resident, which was sometimes confusing, necessitating another round of mailings, with the
appropriate signature line highlighted.
To determine the reasons for not granting consent,
a sampling of families who declined enrollment were sent
anonymous questionnaires and asked to rank several reasons
for refusal to allow their relative to participate in the study.
Participant Rescreening
After obtaining informed consent, each potential participant was rescreened by a research nurse to confirm eligibility.
RECRUITMENT FOR MICRONUTRIENT TRIAL IN ELDERLY PERSONS
Table 2. Response From Proxies Regarding Refusal to Enroll
Table 1. Recruitment Results
Recruited Participants
Total bed capacity of all facilities
Total eligible candidates
Number
%
8026
2814
35%
Female
Male
Competent
Incompetent
2087
727
1152
1662
74*
26*
41*
59*
Total study consents
874
31*
497
376
57#
43#
257
617
166
29#
71#
27y
By patient
By proxy
Rejected at screening
Accepted/enrolled
Dropouts
1471
Notes: From Facilities with total bed capacity of 8026.
*Of eligible candidates.
#
Of those consented.
y
Of those enrolled.
The rescreening process consisted of a chart review, patient
assessment, and height and weight measurements. Blood was
drawn for the measurement of serum albumin. Patients with
albumin levels ,3.0 g/dL were excluded.
Determining Time and Costs for Recruitment
The time to carry out the aforementioned tasks and
associated indirect costs were determined using a retrospective approach. The steps involved in the recruitment process
were identified after enrollment (end of year 3), and each
team member independently estimated the time needed to
complete each step. These estimates were averaged and
multiplied by activity data. The research team then evaluated
the tasks for the purpose of assigning an appropriate
personnel level to each. The following staff levels were thus
identified and assigned to corresponding tasks: level 1, staff
assistant; 2, phlebotomist; 2.5, graduate student; 3, recruiter;
3.5, research nurse; 4, study coordinator; 4.5, recruitment
director; and 5, principal investigator and study physicians.
Salary and fringe benefits were determined for the 3 years
of the project and an average hourly rate was calculated for
each staff level. These rates were multiplied by the total
number of hours used in completing each task. When a task
required a range of staff levels, the three different hourly
rates were averaged based on the participation level of the
different types of study personnel.
Nonpersonnel costs associated with the study included
recruitment materials, postage, office supplies, interpreter
and translator charges, gifts to volunteers, nursing home
staff educational opportunities and gifts, staff travel, and cell
phone expense. The total recruitment cost was calculated by
adding the personnel and supply cost, and then dividing by
617 (the number of enrolled participants) to determine the
enrollment unit cost.
Questionnaires Sent and Returned
Number
%
610
229
36
Questionnaires sent
Questionnaires returned
Reasons for refusal
Relative too old/frail
Objected to blood draws
Objected to skin tests
Additional two pills per day too much burden
Not comfortable with randomization
Disliked medical research in general
Objected to flu vaccine
Objected to outside nurse visits
Relative already taking vitamin E and
refuses to stop for study
61
54
33
24
23
13
12
9
7
Note: Respondents could select more than one option.
with affiliated long-term care programs were contacted, and
all three agreed to participate. Of the 128 independent homes
contacted, 18 (14%) agreed to participate. Thus, a total of 33
facilities enrolled in the study, with 9 homes and 1 institution
participating for more than 1 year. A total of 354 marketing
letters were sent to nursing homes, followed by 592 phone
calls; then 161 site visits were made to participating facilities
for informational, staff training, and consenting purposes.
Recruitment of Potential Participants and
Informed Consent Process
Combined, the cooperating homes provided a total pool
of 8026 residents, of whom only 2814 (35%) were initially
believed to be eligible for the study. There were 1152 visits
to competent patients in their respective nursing home by
the study team resulting in 497 consents. Of the 1622 initial
mailings to families of incompetent patients, only 60
families immediately returned a signed consent; the rest
required an additional 3324 follow-up phone calls to obtain
the additional 316 consents. Table 1 illustrates the result of
these recruitment efforts.
To determine the reasons why proxies refused to enroll
their wards, 610 anonymous questionnaires were sent to
family members who had declined enrollment. They were
asked to report their reasons for study rejection and to
provide additional comments. Two hundred twenty-one
questionnaires (36%) were returned; the results of the
survey are reflected in Table 2.
RESULTS
Role of PCP in the Recruitment Process
Although relatively few PCPs objected to their patient’s
study participation, these objections were not expressed
until their signature was requested on the physician orders.
These objections included an opinion that vitamin E should
be prescribed in larger doses to be effective (or that it was
not effective) or that their patient had medical or psychiatric
contraindications to participation.
Marketing Efforts to Recruit Nursing Homes
Of 20 nursing home corporations contacted, 4 (20%)
expressed interest. A total of 12 homes operated by these
4 corporations participated in the project. Three institutions
Rescreening
The rescreening process resulted in the rejection of 29%
(n ¼ 257) of those who had already provided consent. Table
3 shows reasons for rejection.
GISMONDI ET AL.
1472
Table 3. Reasons for Rejection of Candidates
Reason for Rejection
Table 5. Time and Cost per Recruitment Phase
Number of
Candidates
% of
Candidates
173
7
53
67
3
21
17
7
7
257
3
100
Did not meet eligibility criteria
upon second screen
Medical/surgical
Withdrew consent
Death between time of consent
and enrollment
Discharged between time of
consent and enrollment
Total
Time and Resources
Table 4 illustrates the recruitment staffing requested in the
grant proposal, the staff approved by the grant reviewers,
and the actual staff time required to recruit and enroll the
617 participants. Original projections that the recruitment
tasks could be accomplished by dedicating 10% of the Jean
Mayer/USDA Human Nutrition Research Center on Aging
(HNRCA) recruitment director’s time, a 0.5 full time
equivalent (FTE) recruiter position, and a 0.25 FTE staff
assistant, were inadequate. In addition to the requested
personnel being underestimated, the grant reviewers disapproved funding for the staff assistant position.
Furthermore, it was determined during the study that
some aspects of the recruitment process, such as training of
nursing home staff and consenting competent participants,
was more appropriately performed by the research nurses. A
phlebotomist was hired to draw all screening and study
blood specimens because the research nurse FTE requested
(for aspects of the study other than recruitment) was only
partially funded. In addition, the study coordinator, whose
position and time were not funded by the grant, conducted
all scheduling for screening and enrollment. Thus, the
percent effort for the recruitment director, research nurses,
phlebotomist, and study coordinator for recruitment and
enrollment of the participants had to be increased. These
additions were funded through the HNRCA and supplementary funds from the National Institute on Aging.
The total time and cost for each phase of the recruitment
process is broken down in Table 5. Personnel expenses
accounted for 86% of the total cost, thus each successful
enrollment required an average of 15 hours of staff time at
Table 4. Recruitment Positions Proposed, Funded, and Actual
Staff Position
Recruitment director
Recruiter
Staff assistant
Registered nurse
Phlebotomist
Study coordinator
Graduate students
FTE* (y)
Requested for
Recruitment
0.10
0.50
0.25
0
0
0
0
(3)
(3)
(3)
(3)
(3)
(3)
(3)
Note: FTE ¼ full time equivalent.
FTE* (y)
Approved
by Grant
Review
Committee
0.10
0.50
0
0
0
0
0
(3)
(3)
(3)
(3)
(3)
(3)
(3)
Actual
FTE* (y)
Needed to
Complete
Study
0.40
0.20
0.20
0.50
0.06
0.50
0.05
(3)
(1)
(2)
(3)
(3)
(3)
(3)
Actual
Hours
Needed to
Complete
Study
2184
364
728
2730
328
2730
273
Recruitment Phase
Selection/marketing of facilities
Recruitment of facilities
In-service/identification of
potential study candidates
Notifying primary care providers
Obtaining informed consent
directly from candidates
Obtaining informed consent
from proxies
Scheduling
Screening
Total staff time/cost
Total other cost*
Total cost
Total per enrolled participant**
Total
Time (h)
Staff
Cost ($)
477
57
18,227
4,038
1279
196
42,656
3,781
1229
39,889
1374
2713
1816
9141
43,028
317,661
15
24,459
85,662
55,921
274,633
515
Notes: *Other costs include recruitment materials, postage, office supplies,
interpreter and translator expense, gifts to volunteers, nursing home staff
educational opportunities and gifts, research staff travel, and cell phone charges.
**Total enrolled ¼ 617.
a combined personnel and supply cost of $515 per participant and a total cost of $317,661.
DISCUSSION
During the process of recruiting and enrolling participants
for this multisite micronutrient supplementation trial we
encountered a variety of obstacles. To attain the projected
sample size for this study, a large number of nursing homes
needed to be recruited. To facilitate the process of enrolling
homes into the study, we initially contacted administrators
of nursing home corporations, rather than individual nursing
homes assuming this approach would reduce the time
burden. Approval from these corporate offices facilitated the
process by providing access to numerous facilities, but did
not eliminate the task of marketing to each home. As timing
became critical, we found that direct marketing to individual
homes was more efficient. The targeting of large institutions
maximized the number of potential participants per home,
but required more time in study start-up due to the need for
review and approval by individual Institutional Review
Boards and clearance to access computerized medical
records.
Because we did not anticipate an average of nine PCPs
per home, the importance of effective communication with
PCPs cannot be overly emphasized. The residents and their
families wanted the assurance of knowing that their PCP had
been informed and had approved the patient’s participation.
The most effective way to enlist the PCP’s support proved
to be the letter from the medical director and study
physician, leaving the final decision to them to enroll their
patients in the study. With a 29% rejection rate, the
rescreening process revealed imprecision in the candidate
list developed by the nursing home staff. Despite in-service
training, some floor nurses were unable to select eligible
candidates. Furthermore, their recall of a patient’s medical
conditions, micronutrient supplement consumption, antibiotic use, and influenza vaccine history was not always
RECRUITMENT FOR MICRONUTRIENT TRIAL IN ELDERLY PERSONS
accurate. In general, it would have been preferable for
a study nurse to develop the candidate list from a detailed
record review.
Accurately assessing competency before seeking consent
is a crucial factor in saving time and costs. Without a correct
assessment, the research team’s efforts were doubled,
delaying recruitment by at least 2 weeks (the average time
needed to contact family or health care proxy and to obtain
consent). In addition, the recruiter occasionally experienced
disagreement by the family with the team’s assessment of
the patient’s competency; the family felt that their relative
was capable of signing the consent. When the nursing home
social worker was involved, competency assessments were
more accurate. In addition, the social worker provided
insight into those families who might be more receptive to
the research. It is highly recommended, therefore, that clear
objective criteria such as a paper and pencil test be used
along with a social service assessment to better estimate
competency and willingness, thus avoiding such setbacks
and ambiguity.
Obtaining consent was the most labor intensive, logistically problematic phase of recruitment. The importance of
not burdening nursing home staff with the responsibility of
obtaining informed consent or carrying out study procedures
has been previously noted (3). Furthermore, elderly nursing
home residents may feel inappropriately coerced into study
participation if nursing home personnel are involved in the
consent process (4). We therefore did not ask the homes’
nurses to assist with the informed consent process. However, our experience confirmed the findings of others (3–5)
that cooperation with the nursing home staff was critical
for successful resident recruitment. Their opinions proved
influential in determining resident and family decisions to
participate. It was, therefore, important to carefully prepare
the in-service meeting content and address any negative
perceptions on the part of the nursing staff toward medical
research in general.
Using a recruitment and consent process similar to ours,
Cohen-Mansfield and colleagues (4) reported a consent rate
of 84% for competent residents. We were able to attain only
a 57% rate of consent for competent candidates. Warren and
colleagues (8) found that 54% of health care proxies agree to
consent to research participation for the nursing home
patient they represent. Table 1 demonstrates that our rate of
43% was lower than that reported in the literature. Multiple
blood draws and intradermal skin tests may have dissuaded
participation in both groups of candidates.
Our assumption that daily dispensing of study supplements was not an added burden to the nursing home staff
was not applicable at all participating sites. One nursing
home requested to terminate the study due to an acute
personnel shortage—fortunately, that particular situation
was successfully resolved. In addition, during the course
of this study the Massachusetts Department of Public
Health was citing homes that frequently prescribed more
than eight medications per day per patient. Homes had to
justify the additional supplements for study participants, so
it was important that each was proved with study documentation addressing this issue, should it be raised in an
audit.
1473
Table 6. Recommendations
1. Use all available state government resources, as well as professional and
personal referrals, to identify and select nursing homes.
2. For large clinical trials, select homes with large bed capacities to keep the
number of sites manageable.
3. Long-term care institutions should be explored and recruited at the planning stage of the clinical trial so that all the necessary Institutional Review
Board requirements can be met in a timely fashion.
4. For more effective recruitment efforts, involve the primary care physicians
(PCPs) in the nursing home as early in the process as possible. This not
only helps in the identification of appropriate candidates but also encourages enrollment when the PCP agrees that the study is worthwhile.
5. Perform detailed patient record reviews prior to the consenting process.
6. Provide adequate training sessions and incentives to assure the cooperation of the nursing home staff.
7. Establish objective methods for the determination of mental competency
as part of the protocol and enlist the assistance of the nursing home social
service staff.
8. Anticipate the need for two research team members to be present during
the consenting process.
9. Reduce or eliminate any extra burden on the nursing home staff generated
by the study.
10. Anticipate that state public health regulations pertaining to long-term care
facilities might impede on your study procedures.
11. Use recruitment phases as outlined in Table 4 to estimate project-specific
staff time and costs.
Of the 617 participants enrolled in this study, only 36
(6%) represented minorities. Well documented obstacles to
research participation by ethnically diverse people such as
mistrust of research institutions and language were compounded by the fact that many minority cultures are less
likely to institutionalize their elderly persons. In addition,
we found that many homes lacked documentation of race in
their clinical charts, thus making it difficult to identify
potential minority participants. One nursing home composed of 99% elderly Chinese persons yielded only 8 of 100
potential participants, yet required translated consents and
interpreters to assist the study nurse with weekly collection
of study data. These additional requirements increased both
the time spent by the research nurse and the study costs.
This low yield may have been a reflection of the objection
among elderly Chinese persons to having blood drawn (9).
Gaining acceptance of the study by the nursing home
administration, educating the nursing home personnel,
identifying potential study participants and determining their
competency, obtaining consent from competent candidates
and the health care proxies for incompetent residents, and
performing rescreening procedures all consumed more
personnel resources than had been estimated in preparing
the study budget. The data presented here provide a retrospective assessment of the time and staff costs associated
with recruitment and enrollment of nursing home residents in
this intervention clinical trial. The result of this assessment
indicates that enrollment of each participant required 15
hours of staff time at a cost of $515. The unit time and costs
described above could be modified to meet the differing
requirements of comparable studies.
This information should help investigators to develop
better budget projections for similar types of clinical trials,
as well as grant reviewers and administrators in evaluating
the suitability of the requested budget for such proposals.
1474
GISMONDI ET AL.
In addition, we have put together a list of recommendations that can be found in Table 6.
ACKNOWLEDGMENTS
The following long-term care facilities of Massachusetts are recognized
for their contributions to the Vitamin E and Infection Study: Endicott
Manor of Dedham, Serenity Hill of Wrentham, Meadowbrook of Canton,
Hellenic of Canton, Meadow Green Nursing and Rehabilitation Care of
Waltham, Ellis Nursing Center of Norwood, Laurel Ridge of Jamaica Plain,
Brook Farm of West Roxbury, Mariner at Longwood of Boston, Hebrew
Rehabilitation Center for Aged of Roslindale, Wingate of Sudbury,
Wingate of Brighton, Wingate of Needham, Wingate of Reading,
Sunbridge Care and Rehabilitation of North Reading, Sunbridge Care and
Rehabilitation of Lexington, Sunbridge Care and Rehabilitation of Newton,
Wellesley Nursing and Rehabilitation Center, St. Patrick Manor of
Framingham, St. Joseph Manor of Brockton, Sancta Maria Nursing Facility
of Cambridge, Eastpointe of Chelsea, Eastwood of Dedham, East Village of
Lexington, South Cove Manor of Boston, Hancock Manor of Dorchester,
Regent Park Manor of Brockton, Coyne Healthcare Center of Rockland,
Glen Ridge Nursing Care Center of Malden, Lighthouse Nursing Care
Center of Revere, Kathleen Daniel/Centennial New England of Framingham, Veterans Administration Hospital, Extended Care, Brockton, and
Veterans Administration Hospital, Extended Care, Bedford.
Address correspondence to Simin Nikbin Meydani, DVM, PhD,
Nutritional Immunology Laboratory, Jean Mayer/USDA Human Nutrition
Research Center on Aging Tufts University, 711 Washington St., Boston,
MA 02111. E-mail: [email protected]
REFERENCES
1. Kemper P, Murtaugh CM. Lifetime use of nursing home care. N Engl J
Med. 1991;324:595–600.
2. Gabrel CS. Characteristics of elderly nursing home current residents
and discharges: data from the 1997 National Nursing Home Survey.
Adv Data. 2000;313:1–15.
3. Ouslander JG, Schnelle JF. Research in nursing homes: practical
aspects. J Am Geriatr Soc. 1993;41:182–187.
4. Cohen-Mansfield J, Kerin P, Pawlson G, Lipson S, Holdridge K.
Informed consent for research in a nursing home: processes and issues.
Gerontologist. 1988;28:355–359.
5. Sachs GA, Rhymes J, Cassel CK. Biomedical and behavioral research
in nursing homes: guidelines for ethical investigators. J Am Geriatr
Soc. 1993;41:771–777.
6. Meydani SN, Leka LS, Fine BC, et al. Vitamin E and respiratory tract
infections in elderly nursing home residents: a randomized controlled
trial. JAMA. 2004;292:828–836.
7. Division of Health Care Finance and Policy, Commonwealth of
Massachusetts. Massachusetts Nursing Home Utilization: Report Card.
Boston, MA: Executive Office for Health & Human Services; 1993.
8. Warren JW, Sorbal J, Tenney JH, et al. Informed consent by proxy.
N Engl J Med. 1986;315:1124–1128.
9. Holroyd E, Molassiotis A. Hong Kong Chinese perceptions of the experience of unrelated bone marrow donation. Soc Sci Med. 2000;51:
29–40.
Received August 4, 2004
Accepted October 14, 2004
Decision Editor: John E. Morley, MB, BCh