Printer Friendly

Program factors that influence utilization of adult day care.

Adult day care (ADC) is a community-based group program designed to meet the needs of functionally impaired adults through an individualized plan of care. It is a structured, comprehensive program that provides a variety of health, social, and related support services to groups of adult clients in a protective setting. Adult day care is increasingly viewed as a promising long-term care option because it allows spouses, siblings, and adult children of impaired older adults regular and reliable respite from caregiving.

The field of ADC is expanding very quickly, from 18 centers in 1974 (Weissert 1977) to over 1,700 today (Von Behren 1988). Despite this rapid growth, a recent national survey found that ADC centers are somewhat underused and underattended (Conrad, Hanrahan, and Hughes 1990). Specifically, the 1986 Adult Day Care Assessment Procedure (ADCAP) survey of ADC programs in the United States obtained detailed information about the structure, client population, and process of ADC centers. This national census survey of 974 ADC centers found that only 29 percent had a waiting list and that the ADC centers, on average, fell short of daily capacity by four clients for a nonattendance rate of about 18 percent. If ADC is to be a viable, cost-effective long-term care option, ADC centers will need to attract and maintain clients efficiently. To control costs, it will be important to utilize daily capacity by keeping attendance rates as high as possible. While previous research has focused principally on macroenvironmental factors, this article uses data from the ADCAP national survey to examine the strength of the influence of microenvironmental characteristics on ADC center utilization. If ADC centers are to play a role in improving the cost effectiveness of long-term care, it is essential that we learn how their services affect demand and attendance rates.


One useful way of studying utilization is to examine two types of facilitating factors, namely, the macroenvironment and the microenvironment. Macroenvironmental factors refer to the characteristics of the setting in which the program is located and include geography and climate, cultural factors, economic and political factors, demographics, legal constraints, and community organization (Moos 1974). Microenvironmental factors are the characteristics of the health care resources and services that are brought to bear to address the wants and needs of consumers. These include the built environment, the philosophies and abilities of administration and staff, the social milieu, activities, and the characteristics of the clients who actually use the service (Conrad and Miller 1987; Conrad and Roberts-Gray 1988).

Several macroenvironmental factors have been found to influence health care utilization. Distance and the interaction of distance with sociodemographic characteristics such as age, sex, race, occupation, and degree of mobility have been found to affect contacts with physicians, hospitals, and other health care resources (DeVise 1968; Weiss, Greenlick, and Jones 1971; Fuller 1974; McGuirk and Porell 1984; Shannon, Skinner, and Bashur 1973; Haynes and Bentham 1982; Joseph and Phillips 1984; Gesler and Meade 1988). In examining interactions among distance, sociodemographic, locational, and activity-space factors, Gesler and Meade (1988) found that urban ecological structure, distance, and degree of mobility appeared to play a more important role in health care utilization than did the personal characteristics of individuals in a small metropolitan area.

Regarding ADC in particular, it is likely that the age structure of the community (macro) would influence the need for ADC and would affect utilization. Economic constraints are also important macroenvironmental factors for new programs like ADC. At present, ADC is not reimbursed by Medicare. Some states support ADC via Medicaid waivers while others do not. Specifically, Medicaid made up the largest portion of the budgets of ADC centers in only seven states (Von Behren 1988). Much of ADC is supported through client fees and private or public organizations such as churches and hospital systems. In 17 states, participant fees accounted for the largest portion of the ADC centers' budgets (Von Behren 1988). Therefore, state variability in ADC funding mechanisms and levels is likely to have a bearing on utilization.

Several theories emphasize the importance of the characteristics of health care consumers in determining utilization. While consumer characteristics have previously been treated separately from macro- and microenvironmental factors, some consumer characteristics are appropriately subsumed under macro, others under micro. Therefore, the community sociodemographic context -- a macroenvironmental component -- includes consumer demographic variables, whereas functional status of the client population being served in the ADC center is subsumed under the microenvironment; that is, it is a characteristic of the ADC center itself that may influence utilization. In the case of ADC, the consumer is usually a dyad of the client and the informal caregiver who is usually a relative. Although informal caregivers are recognized theoretically as a significant influence, data about them were not available for the analysis presented here.


One of the earliest theories addressing health care utilization, the Health Belief Model (Rosenstock 1966; Becker and Maiman 1975), focused on perceptions of consumers regarding susceptibility to illness, seriousness of illness, barriers to health care access, benefits, and cues to action and their relationship to engagement in health-related behaviors, including the utilization of health services. Suchman (1967) emphasized the influence of social groupings and linkages. Andersen's model (1968) combined the previous models while focusing on family composition, social structures, and health beliefs. Gross (1972) put this basic model in the form of an equation: U = f|E;P;A;H;X~ + e, where U = utilization; f = a function of; E = enabling factors (e.g., income); P = predisposing factors (e.g., attitudes); A = accessibility factors (e.g., distance); H = perceived health level; X = individual and areawide exogenous variables (e.g., age, sex, race); and e = residual error term.

While the foregoing models all built upon one another, they omitted a conspicuous factor in the explanation of health services utilization -- the characteristics of the service delivery system. In 1974, Aday and Andersen proposed a comprehensive framework for the study of access to health services that included macroenvironmental factors, consumer factors, and the characteristics of delivery systems (micro), the organization, quality, convenience, cost, coordination, and courtesy of the latter. Concurrently, Donabedian (1973) described the characteristics affecting access as geographic (macro) and socio-organizational (micro). Socio-organizational accessibility included all the attributes of the service itself, such as cost, intake policies, specialization, hours of operation, provision of transportation, and marketing strategies. Moos (1974) described and emphasized the importance of the influence of the social environment within each program (micro) on the response of clients to the service delivery system.


This article applies the foregoing theoretical work on utilization to the rapidly growing program of ADC, and expands upon service utilization theory by testing the specific contribution of the characteristics of the service delivery systems, including perceived quality of services, to health care utilization. Previous research has not tested the influence of microenvironmental factors principally because good measures of program microenvironments have not existed (Conrad and Roberts-Gray 1988).

The general theory that guides the present study is depicted in Figure 1. This theory is consistent with the literature cited in the previous section as well as with more recent explanations (Ricketts et al. 1987) of health care utilization. In this theory, both center demand and attendance rate are affected by macroenvironmental factors that include state policies and regulations, funding resources, community sociodemographic context, and community health care resources; and by microenvironmental factors that include characteristics of structure (including facility, administration and staff, and characteristics of the client population and the clients' informal caregivers) and process (including characteristics of the social environment and of the services and activities).

While reciprocal relationships are logical, their specification and testing would greatly increase the complexity of the model and the analysis, and would draw upon another body of literature with much less theoretical development. The scope of this article is limited to an attempt to achieve greater clarity about the influence of macro- and microenvironmental factors on utilization. Therefore, it does not examine the influence of utilization on the macro- and microenvironments, although this clearly is an important next step.

The figure does not depict the relative influences of demand versus attendance. It is reasonable to expect, however, that macroenvironmental factors will have a stronger influence on demand whereas microenvironmental factors will have a greater effect on attendance rate. For example, we can see clearly how urban/rural setting (macro) would have a strong influence on demand, but it is less clear how it would affect attendance rate. On the other hand, it is logical that the characteristics of transportation services (micro) of a center could influence clients to stay at home and would be strongly related to attendance rate.


The data were obtained from a 1986 national census survey of ADC in the United States using the Adult Day Care Assessment Procedure (Conrad and Hughes 1989; Conrad, Hanrahan, and Hughes 1990). The principal source for the sample was the entire population of ADCs listed in the updated directory of ADCs compiled by the National Institute on Adult Daycare (Robins 1985). This directory was further updated during the survey by mail and telephone contacts with providers regarding new centers and centers that had closed. Of 1,347 centers, 974 or 72.3 percent provided usable responses, a high rate for a dense, 24-page mailed questionnaire. The questionnaire was directed to the on-site director of each center, who was assumed to be the single respondent with the most complete, direct, and intimate knowledge of all phases of the program. To ensure true and complete responses, confidentiality was assured. It should be recognized that the individual administrator's perspective might not have coincided with the perspectives of other stakeholders, such as the client, home caregiver, or other staff. However, direction of bias was not clear since some administrators tended to overrate while others underrated their centers.


Center Demand. Center demand was defined for this study by the number of clients enrolled at the center plus the number on the waiting list. The number enrolled included all clients on the ADC roster whether they were attending five days a week or only one day a week. The reason for summing the number of clients enrolled in the program with the number on the waiting list was to obtain an indication of how much the services of a particular ADC center were sought after. The waiting list was included because it would indicate those centers where demand exceeded the ability to provide services. This situation was viewed as favorable since such ADCs would be fulfilling a perceived need and would have a greater chance of surviving, growing, and being successful. The factors influencing demand were of great interest to ADC providers and policymakers since they might indicate locations where ADC would flourish (macroenvironmental factors) as well as the ADC characteristics that would promote improved center demand (microenvironmental factors).

Attendance. Attendance rate was defined as the average weekday attendance rate per center (average number of clients attending the center weekdays divided by the average number enrolled weekdays). The average number attending is simply the number of clients who participate in ADC on the average weekday. The average number enrolled is how many people were expected to attend on the average weekday. Therefore, an ADC center that expected 20 clients to attend each day, but actually saw only 15, had an attendance rate of .75. The factors influencing attendance rate are important to ADC providers because they want to utilize their resources efficiently to full capacity in order to maximize funding and remain viable. The attendance rate is conceived of as being influenced by the need for or dependence on the service by clients and informal caregivers. This can also be seen as a reflection of the center's ability to satisfy client and informal caregiver needs for such things as socialization, transportation, day-to-day caregiver respite, client health care, and supervision.

Unit of Analysis. The principal contrast between these definitions of ADC center demand and attendance rate and those of other organizational and market paradigms is that we use the center as the unit of analysis, whereas most other studies have operationalized utilization with the individual or the event as the unit of analysis (Beland 1988). The center is the more appropriate unit when studying group-delivered programs (Whiting-O'Keefe, Henke, and Simborg 1984; Koepke and Flay 1989).


The independent variables were measured with the Adult Day Care Assessment Procedure (ADCAP). The ADCAP is a set of scales that measure three major components of ADC programs: structure, process and client characteristics (Conrad and Hughes 1989). Tables 1 and 2 present the scales, Cronbach's alpha reliability estimates, zero-order correlations with enrollment and attendance, and sample items of the ADCAP.

Macroenvironment. Although macroenvironmental factors were not a major focus of this study, we attempted to control for these by including a measure of urban versus rural character, that is, whether the center was or was not located in a Metropolitan Statistical Area (MSA), and whether or not the center was licensed. Additionally, the total number of linkages for referral to and from 16 relevant outside agencies indicated the richness of the referral network available (White 1979; Shortell 1972). We did not have reliable or complete information on funding resources nor on community demographics.

Microenvironment: Structure. Structure refers to the center resources (e.g., staff, built environment, technology) that enable service provision. The structural features that are likely to influence enrollment and attendance at the facility include three general categories: (1) accessibility, (2) facilities, and (3) client population.

Several ADCAP variables were judged to be indicators of accessibility. Promoting ADC Awareness assessed the degree to which the center advertised and educated people about its services throughout the community. Provision of Transportation counted the number of vehicles and their wheelchair accommodations. We also measured the percentage of clients for whom the center provided transportation. The age of the center was an indicator of how much time it had had to develop a clientele. The average hours open on weekdays was thought to promote access especially to caregivers employed outside the home.

The ADCAP Inservice scale assessed the degree of provision of inservice staff training. Pleasantness assessed characteristics important for enjoyment and comfort in the center. Amenities for the Disabled assessed characteristics important for the comfort of disabled persons. Space measured square footage and number of rooms. Safety assessed the provision of safety features.

In this analysis, the client population was regarded as a microenvironmental characteristic. In other words, a center may have a mission to serve a certain type or types of client; in turn, the type of client population targeted may influence the size and draw that the center has. The Activities of Daily Living and Instrumental Activities of Daily Living ADCAP scales assessed the functional ability of the population in each center. To obtain an indication of the level of cognitive impairment in the client population, we also used the percentage of clients with Alzheimer's disease.



Microenvironment. Process. Process was measured by the Services and Activities and the Social Environment scales. Using the first scale administrators rated whether a set of activities and services occurred: (1) rarely, never, or on admission only; (2) a few times a year; (3) more than twice a month; (4) weekly; and (5) daily.

Clinical Services assessed the provision of services such as hearing examinations and dental care. Care Planning measured the provision of several activities important to the planning and management of client care. Counseling assessed a variety of counseling and psychosocial support services. Personal Care and Training measured the provision of low technology training, exercise, and rehabilitation services. PT, ST, OT assessed provision of formal physical therapy, speech therapy, and occupational therapy. Therapeutic Recreation measured a set of formal recreational activities designed to provide therapeutic benefits. Entertainment was a set of social activities designed for the clients' enjoyment. Client Education involved educational programs for clients. Support for Families included educational and supportive programming for the family members of ADC clients. Help with Self-Care measured activities to support clients with heavy care needs. Pastimes assessed involvement in crafts, games, and hobbies.

In the second set of Social Environment scales, administrators rated eight characteristics of the social environment of the centers. Staff Cooperation assessed the degree to which staff were perceived as having good working relations. Morale Problems indicated problems in morale, primarily in clients. Independence Promotion assessed the degree to which clients initiated or were encouraged to initiate activities. Communication indicated the degree to which client-staff interaction was perceived as warm and attentive. Director Control measured degree of centralization and director control. Social Involvement measured the importance placed on involvement of family and community.


Because over 150 items were involved in the composition of the scales used in these analyses, it was impractical to use listwise deletion of missing data, that is, dropping a case with any missing data. We wanted to include those items with most of their data intact as well as those cases having valid responses for most of the items in the analyses. To this end, any variable included in the analyses could have no more than 30 percent of its data missing. No case was included that had missing data on more than 10 percent of the items. This rule resulted in 822 cases. For these cases the grand mean was substituted for any missing value. The distributions of all variables were examined for conformity to the assumptions of regression analysis. A skewness of plus or minus 1.5 was used as the normality criterion. Four variables were skewed beyond |+ or -~ 1.5. After the appropriate transformations, they also met the normality criterion.

Our major interest in this study was to test the hypothesis that five sets of variables would have a statistically significant relationship to center demand and attendance rate, and that multiple |R.sup.2~ would be statistically significant. In addition to the confirmatory analyses we examined the effect of particular variables within the sets in order to refine our theory and propose directions for future research. To accomplish these goals, we analyzed the data in three steps. First, we obtained the simple correlations of independent variables with the demand and attendance indicators. The underlying assumption was that each variable was chosen because, theoretically, it should correlate with the dependent variables. If it did not, it failed this hypothesis test and was dropped from further analyses. The second analysis examined the influence of each set when all other sets were also in the equation. This would indicate which sets had the largest independent influence controlling for all other sets. The third analysis examined the influence of each independent variable when all of the others were also in the equation. This analysis identified those variables with the largest independent influence on the dependent variables.


Macroenvironment. Since enrollment was operationalized as the number enrolled plus the number on the waiting list, a major macroenvironmental influence was location in a Metropolitan Statistical Area (MSA), a common indicator of urban community. Large urban communities are more densely populated and have a greater number of eligible clients within a convenient distance. Another macroenvironmental influence was the availability of other health care resources to serve as a referral network. This construct was operationalized as the number of referral linkages that the ADC center had with 16 other types of health care providers. Therefore, location in an MSA and the number of linkages were entered as a partial control for factors external to the ADC centers. We recognized that the number of linkages was also determined in part by the ADC center itself, but decided that the major influence on linkages was the richness of the health care referral network.

Microenvironment: Client Population. We hypothesized that the second strongest influence would be the characteristics of the client population, a microenvironmental characteristic. Therefore, the functional and cognitive status ADCAP measures were included.

Social Environment, Services and Activities, and Structure. We hypothesized that process factors such as the social environment (e.g., morale) of the center and the availability of services and activities would create a therapeutic milieu that would influence retention of clients and would foster a program reputation attractive to prospective clients. Therefore, we hypothesized that the Social Environment scales would have the third strongest influence and the Services and Activities scales fourth. Although structural factors were considered important, we hypothesized they would be slightly less important than the preceding variables because variables such as safety factors, pleasantness, and staff training would affect the process scores and share the same variance.

In summary, these expectations stated that community size and number of linkages would explain the largest amount of variance in demand. The second largest amount would be explained by client population characteristics. Next the process factors of social environment and delivery of services and activities would have the third and fourth largest influence, respectively. Finally, the structure factors such as staff training and pleasantness would have a significant influence on enrollment but smaller than the preceding variables.


In the analysis of attendance rates, the same sets of variables were entered but with different expectations. Community size and number of linkages were expected to be significant. Client population characteristics would be significant. Process factors would have more influence on attendance than on demand. Services and activities would have a significant influence. Finally, some structural influences were still expected to be quite strong, especially transportation; and, overall, structural characteristics were expected to be statistically significant.


The results provide previously unavailable information, obtained from the on-site administrators of the ADC centers, about community factors, the client population, social environment, services and activities, and structure of ADC. These results either support or bring into question the influence of factors that we expected to be associated with utilization as operationalized by demand and attendance in ADC.


Since the variables were chosen on the assumption that they would correlate with demand, those without a significant correlation at the liberal criterion of p |is less than~ .1 clearly did not meet the assumption and therefore were dropped from further analyses. Twenty-five of the 30 variables were significantly correlated and were judged useful in the subsequent regression analyses. In the regression analysis the set of three macroenvironmental variables -- Located in SMSA, Total Number of Linkages, and Licensure -- explained a significant percentage of variance. A set of client population descriptors -- Percentage of Alzheimer's Clients, Sum of ADL and IADL Scores -- was also significant although not as strong as expected.

Unexpectedly, the set of social environment scales (Independence-Promoting Approach, Morale Problems, Communication, and Director Control) was found not to explain a significant percentage of variance. However, the set of ten Services and Activities scales did explain a significant additional percentage of variance. The most influential was the structure set. The variance explained by the full model was 22 percent.

The examination of the t-values of each variable with all other variables in the equation provided evidence about those components that had the greatest independent influence on demand. The three community variables were significant. Centers located in urban areas and those with more referral linkages with community agencies and licensure had higher demand.

Of the two client population variables, the combined ADL/IADL score was significant, indicating that centers serving higher- functioning clients had higher demand. Conversely, the results suggest that centers serving clients with poorer physical and cognitive functioning had lower demand.

The individual Social Environment scales did not have a significant association with demand, controlling for all other variables in the equation. The service variable that was most strongly associated with greater demand was Client Education, which consisted of two items -- educational programs for clients and classes or lectures.

The structural characteristics most strongly associated with TABULAR DATA OMITTED demand were the number of vehicles available for use, auspices (-), provision for more staff in- service training, safety features (-), and being incorporated (-).


The correlation analysis indicated that few of the hypothesized relationships between independent variables and attendance were actually significant at the liberal criterion of p |is less than~ .1. However, at least one variable from each of the five general categories met this criterion and was used in the regression analysis.

In the regression analysis, Location in MSA was not significant. The Functional Status Total and the social environment variable, Independence-Promoting Approach, were also not significant. The set of three Services and Activities scales did have a significant influence on attendance as did the last set of structural variables.

An examination of the t-values for the variables in the full model TABULAR DATA OMITTED indicated that the services and activities, individually, did not have a significant influence; but working together, the factors that were associated with increased attendance were (1) therapeutic services (PT, ST, OT); (2) personal care activities such as help with toileting and grooming, training in ADL, and therapeutic recreation; and (3) a lesser degree of support for families such as providing family support groups and educational programs. The structural features most strongly and negatively associated with attendance were provision of staff training and the pleasantness of the center environment (general attractiveness, cleanliness, lighting, odors, window areas). Positively associated structural factors were the number of vehicles used by the center, and the accessibility and safety features (i.e., no raised thresholds, wheelchair accommodations).


Because this study is an early foray into a very complex area, the findings should be regarded as a suggestive exploration, not a definitive test of a well-specified model. In general, these results do support existing theories; in addition, they provide insight into some factors that may influence demand for and attendance in ADC centers.

The findings indicate that location of an ADC center in an MSA (i.e., urban) environment results in higher demand but not a higher attendance rate. Client functional status, both physical and cognitive, influences demand. Higher functioning is associated with higher demand but not with a higher attendance rate. Clearly, the number of high-functioning clients that could be supported by a center is larger than the number of low-functioning clients that would usually be supported by a similar center.

The results did not support an effect of the social environment of an ADC center on demand or attendance. However, the validity of this finding is more questionable than the other findings because the evaluation of the social environment was made by the ADC administrator. The client perspective, which would have a much greater influence on center demand and attendance, should be obtained in future studies. Further, the home caregivers play important roles that probably influence both center demand and attendance rate.

Services and activities did influence both demand and attendance rate. The services that had the strongest effects on demand were provision of clinical services such as hearing examinations, dental care, and optometry services; therapeutic recreation (i.e., music and art); and client education programs, classes, and lectures. Of course, all of these services may be services provided by the larger centers, which have more support for additional services.

Together, PT, ST, OT, and personal care services (i.e., help with toileting and grooming, training in ADL and IADL) were associated positively with attendance rate, while support for families was negatively associated. These findings reflect greater attendance rates in centers for clients with heavy care needs. These may be rehabilitation model ADCs since they do not provide support (groups, education) for families, as would be needed for clients with chronic long-term health problems.

Structural features have a strong influence on center demand and attendance rate. Positive associations with demand were observed for in-service training and the number of vehicles used by the center. Conversely, safety was negatively associated. This is consistent with the belief that low- functioning clients require more safety features and smaller centers. The negative findings for Auspices and Incorporated indicate that privately funded centers and centers that were incorporated had lower center demand than did publicly funded centers. This finding of lower demand in privately funded centers helps to explain why the survey found that only 10 percent of ADC centers were privately funded (Conrad, Hanrahan, and Hughes 1990).

In contrast, more staff in-service and center pleasantness were associated with lower attendance rates, whereas the number of vehicles and safety features were nonsignificant. Having more in-service sessions and a more pleasant environment may be a function of being larger. In a post hoc analysis, we observed that greater center demand was negatively associated with attendance rate (p |is less than~ .0001). The larger centers were observed to have more staff, more hours of in-service, and a more pleasant environment with a less impaired clientele. These findings suggest that the attendance rates in these larger centers may be lower because the higher functioning clients have a greater choice about attending as well as a higher number of alternative activities. Since these findings are weak, more theoretical and empirical work is needed to explain attendance rates in ADC centers.

What does it mean that greater impairment is associated with lower demand? In health services generally, poorer health is associated with greater utilization (Muller 1986). In the field of ADC this is also the case, since ADC continues to grow along with our growing population of frail older persons. In most studies of health services utilization, either the number of individuals receiving the service or the event/occurrence of each usage is the unit of analysis (Beland 1988). This is because most health care is delivered on an individual basis with distinct episodes or units of care. In our study, utilization was measured with the ADC center as the unit of analysis. This strategy is recommended as more appropriate for studying group-delivered programs (Whiting-O'Keefe, Henke, Simborg 1984; Koepke and Flay 1989). When we examine demand on a per center basis, we find that poorer health is associated with lower center demand.

Therefore, this nontraditional form of health service should not be thought of as being similar to health services that have overwhelmingly been delivered on an individual basis, but rather as showing a similarity to other group-delivered services such as educational programs. For example, in classrooms, the greater the student impairment, the smaller the class size (Forness and Kavale 1985).

Our study indicates evidence of this for ADC, also, that may generalize to health promotion and education programs. Therefore, it is likely that as ADC centers grow, the individual center will reach a certain size, and a new site or additional rooms will then be opened to accommodate the excess demand. Like classrooms, ADC centers may choose to specialize based on the needs of certain types of clients (Cole, Vandercook, and Rynders 1988) such as those with Alzheimer's disease or those in transition from acute care who require rehabilitation.

Rural ADC centers tend to be smaller. Therefore, like rural classrooms, they would be less able to group participants by need but would tend to be more general-purpose in nature. In this regard, a much lower percentage of Alzheimer's clients were found to be served by rural versus urban ADC centers (Conrad et al. 1991), indicating less ability in rural settings to accommodate the special needs of this population.


This study was limited in that it did not specify the direction of influence a priori. For instance: Does providing more clinical services cause higher demand or does higher demand affect the provision of more clinical services? While current theory acknowledges that reciprocal relationships exist, the specific nature of these relationships remains vague. While the scope of this study was limited to theory-testing regarding the factors that influence utilization, future work should develop and test hypotheses about the influences of utilization on macro- and microenvironmental factors.

In the demand analysis, the independent variables explained 22 percent of the variance. What accounts for the remaining 78 percent? The plausible factors include random variation (i.e., chance); factors that have been left out of the model (i.e., specification error); measurement error; and respondent bias. Random variation is self-explanatory. Regarding specification error, it is obvious that important macroenvironmental and microenvironmental factors were not addressed adequately in this study. These are discussed further in the following section. Regarding measurement, it is clear from our presentation of the alphas of our measures that no variables were perfectly measured. This means that all of the correlations were actually slight underestimates of the associations.

Attitudinal factors such as health beliefs were not well measured. Respondent bias is a plausible reason for the small associations derived for the Social Environment scales. Specifically, the administrator's perception of the physical and social environment and the service may not reflect the consumer's perspective adequately; and it is the consumer's perspective, that of both client and home caregiver, that is crucial to demand and attendance.


This study found several program characteristics that appear to affect ADC center demand and attendance rate. To improve the explanatory power of this type of analysis, several additional variables should be included in future studies.

Center Demand. Macroenvironmental variables appeared to have a strong effect and were not well measured in this study. These include state policies and regulations, community sociodemographic factors, funding resources, and community and health care resources. While this study demonstrated the strong influence of microenvironmental factors such as client characteristics, services and activities, and structure, it is likely that the social environment also has a strong influence but that it should be measured with the client as the informant. Additionally, the influence of the informal caregiver is a major factor which should be included. Health beliefs have been found to influence demand and should be assessed for both clients and their informal caregivers. Staff characteristics and assessments of programs are other untapped but potentially useful variables.

Attendance Rate. Two sets of variables, services and activities and structural features, explained small, but statistically significant proportions of variance in ADC attendance rates. Macroenvironmental characteristics were not significant in this study, but may have some influence and should be included in future studies. It is likely that the major reason for the weak findings was that the administrator was a poor informant for variables that should be measured on clients and their informal caregivers. These include client impairment levels; client rating of the social environment, services, activities, and structural characteristics; and client health beliefs. Caregivers should be assessed on objective and subjective burdens of caregiving and on their ratings of those ADC program characteristics likely to affect them such as transportation, hours of service, cost, and supportive and educational programming for them.

In summary, the strategy of obtaining center level data is useful in understanding the factors that influence ADC center demand and attendance rates. To improve our understanding of factors influencing demand, more and better macroenvironmental and client and informal caregiver level measures must be included and aggregated, with the center as the unit of analysis (Whiting-O'Keefe, Henke, and Simborg 1984; Koepke and Flay 1989). To explain attendance rates better, both macro- and microenvironmental variables are needed; but client and informal caregiver (i.e., consumer) factors are likely to be most influential.


The authors are grateful to Michelle Nesbitt for assistance in preparing the manuscript.


Aday, L. A., and R. Andersen. "A Framework for the Study of Access to Medical Care." Health Services Research 9, no. 3 (1974): 208-220.

Andersen, R. A Behavioral Model of Families Use of Health Services. Chicago: University of Chicago, Center for Health Administration Studies, 1968.

Becker, M. H., and L. A. Maiman. "Sociobehavioral Determinants of Adherence with Health and Medical Care Recommendations." Medical Care 13, no. 2 (1975): 10-21.

Beland, F. "Utilization of Health Services as Events: An Exploratory Study." Health Services Research 23, no. 2 (June 1988): 295-310.

Cole, D. A., T. Vandercook, and J. Rynders. "Comparison of Two Peer Interaction Programs: Children with and without Severe Disabilities." American Educational Research Journal 25, no. 3 (1988): 415-39.

Conrad, K. J., P. Hanrahan, and S. L. Hughes. "Survey of Adult Day Care in the U.S.: National and Regional Findings." Research on Aging 12, no. 1 (1990): 36-57.

Conrad, K. J., and T. Q. Miller. "Measuring and Testing Program Philosophy." In Using Program Theory in Evaluation. Edited by L. Bickman. San Francisco: Jossey-Bass, 1987.

Conrad, K. J., and C. Roberts-Gray. "Editors' Notes." In Evaluating Program Environments. Edited by K. J. Conrad and C. Roberts-Gray. San Francisco: Jossey-Bass, 1988.

Conrad, K. J., and S. L. Hughes. Adult Day Care Assessment Procedure.' Research Manual. Northwestern University, Center for Health Services and Policy Research. Working paper #188, 1989.

Conrad, K. J., C. I. Hultman, S. L. Hughes, and P. Hanrahan. "Rural/Urban Differences in Adult Day Care." Northwestern University, Center for Health Services and Policy Research. Working paper #225, 1991.

De Vise, P. "Methods and Concepts of an Interdisciplinary Regional Hospital Study." Health Services Research 3, no. 3 (Fall 1968): 166-73.

Donabedian, A. Aspects of Medical Care Administration. Cambridge, MA: Harvard University Press, 1973.

Forness, S. R., and K. A. Kavale. "Effects of Class Size on Attention, Communication and Disruption of Mildly Mentally Retarded Children." American Educational Research Journal 22, no. 3 (1985): 403-12.

Fuller, G. "On the Spatial Diffusion of Fertility Decline: The Distance-to-Clinic Variable in a Chilean Community." Economic Geography 50, no. 4 (1974): 324-32.

Gesler, W. M., and M. S. Meade. "Locational and Population Factors in Health Care-seeking Behavior in Savannah, Georgia." Health Services Research 23, no. 3 (August 1988): 443-62.

Gross, P. F. "Urban Health Disorders, Spatial Analysis and the Economics of Health Facility Location." International Journal of Health Services 2, no. 1 (1972): 63-84.

Haynes, R. M., and C. G. Bentham. "The Effects of Accessibility on General Practitioner Consultations, Out-Patient Attendances and In-Patient Admissions in Norfolk, England." Social Science and Medicine 16, no. 5 (1982): 561-69.

Joseph, A., and D. Phillips. Accessibility and Utilization: Geographical Perspectives on Health Care Delivery. New York: Harper & Row, 1984.

Koepke, D., and B. Flay. "Levels of Analysis." In Evaluating Health Promotion Programs. Edited by M. T. Braverman. San Francisco: Jossey-Bass, 1989.

McGuirk, M. A., and F. W. Porell. "Spatial Patterns of Hospital Utilization: The Impact of Distance and Time." Inquiry 21, no. 1 (1984): 84-95.

Moos, R. H. Evaluating Treatment Environments: A Social Ecological Approach. New York: John Wiley & Sons, Inc., 1974.

Muller, C. "Review of Twenty Years of Research on Medical Care Utilization." Health Services Research 21, no. 2, part 1 (June 1986): 129-44.

Ricketts, T. C., T. R. Konrad, J. S. Stein, and G. H. DeFriese. "Population Ecology and Health Policy Analysis: The Case of Rural Primary Care Centers." Medical Care Review 44, no. 2 (1987): 345-73.

Robins, E. G. Directory of Adult Day Care Centers. Baltimore, MD: Health Care Financing Administration, Department of Health and Human Services, 1985.

Rosenstock, I. M. "Why People Use Health Services." Milbank Memorial Fund Quarterly 44 (1966 supplement): 94-127.

Shannon, G. W., J. L . Skinner, and R. L. Bashur. "Time and Distance: The Journey for Medical Care." International Journal of Health Services 3, no. 2 (1973): 237-44.

Shortell, S. M. A Model of Physician Referral Behavior: A Test of Exchange Theory in Medical Practice. Chicago: University of Chicago, Center for Health Administration Studies, 1972.

Suchman, E. A. Evaluative Research: Principles and Practice in Public Service and Social Action Programs. New York: Russell Sage Foundation, 1967.

Weiss, J. E., M. R. Greenlick, and J. F. Jones. "Determinants of Medical Care Utilization: The Impact of Spatial Factors ." Inquiry 8, no. 1 (1971): 50-57.

Weissert, W. G. "Adult Day Care Programs in the United States. Current Research Projects and a Survey of 10 Centers." Public Health Reports 92, no. 1 (1977): 49-56.

White, A. N. "Accessibility and Public Facility Location." Economic Geography 55, no. 1 (1979): 18-35.

Whiting-O'Keefe, Q. E., C. Henke, and D. W. Simborg. "Choosing the Correct Unit of Analysis in Medical Care Experiments." Medical Care 22, no. 12 (1984): 1101-14.

Von Behren, R. Adult Day Care: A Program of Services for the Functionally Impaired. Washington, DC: The National Council on the Aging, Inc., 1988.
COPYRIGHT 1992 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Conrad, Kendon J.; Hughes, Susan .; Wang, Shenglin
Publication:Health Services Research
Date:Oct 1, 1992
Previous Article:Home care for the disabled elderly: predictors and expected costs.
Next Article:Mammography referrals for elderly women: is medicare reimbursement likely to make a difference?

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters