Social work Admission Assessment Tool for identifying patients in need of comprehensive social work evaluation.
Hospital social workers play an integral role in providing adequate posthospital care for patients (Fillit et al., 1992; Holliman, Dziegielewski, & Datta, 2001). However, in light of rising health care costs, many hospitals are under fiscal constraints, which creates more pressure for social workers to discharge patients more efficiently and expeditiously (Dobrof, 1991). A physician-based process by which referrals are made only after the medical issues have been addressed and the patient is "ready to go home" is a common but inefficient practice. In such cases, the social work evaluation begins late in the course of the hospitalization, there is increased pressure to discharge the patient, and the full range of services that can be offered in such a short time frame may be limited.
The Social Work Admission Assessment Tool (SWAAT) was developed to identify patients who have complicated discharge-planning needs and require early social work evaluation. The design and conceptualization of this tool was a result of collaborative efforts undertaken by social workers and physicians in the Department of Internal Medicine. Implicit in the development of this tool is the idea that effective discharge planning should begin on admission (Tennier, 1997; Wachtel, Fulton, & Goldfarg, 1987). Identifying patients who require a social work evaluation earlier may provide more time for the social worker to get to know the patient and his or her family, to thoroughly assess the needs, and to coordinate the postdischarge services.
Several approaches to identifying patients at high risk of becoming disposition problems have been developed. Berkman and colleagues (1980) developed the High Social Risk form (HSR) based on the patient's severity of illness, the presence of chronic illnesses, and illnesses that are physically disabling. Glass and Weiner (1976) developed indicators based on the patient's continence, ambulation, age, social background, and thought process (CAAST). Glass and Weiner also developed the 4-Score, a tool designed to predict a patient's nonmedical days based on age, change in residence at the time of discharge, disorientation, and duration of disorientation. These measures are brief and were developed on the basis of characteristics that can be obtained at admission, but it is unclear whether the outcomes were obtained blindly. More recently, the Complexity Prediction Index (COMPRI) was developed to detect patients with complex needs (Huyse et al., 2001). It relied primarily on the clinicians' perception of what a patient needs and did not include an assessment of the patient's perception of need.
In developing the SWAAT, we build on the strengths of existing instruments. Our objectives were three-fold: (1) to develop a reliable and valid screening instrument to identify patients likely to need a social work evaluation; (2) to develop an instrument that was comprehensive and addressed different dimensions of patients' medical, functional, and social needs; and (3) to develop an instrument that was brief enough for practical use and could be used to screen patients within 24 hours of admission.
To generate items for the SWAAT, we evaluated items currently being used by the social workers at our institution to designate high-risk patients. In addition, we searched the literature for established instruments on health and functional status such as the Medical Outcomes Study Short Form 12 (SF-12) (Ware, Kosinski, & Keller, 1996a, b), Katz Activities of Daily Living Instrument (ADL) (Katz, Ford, Moskowitz, Jackson, & Jaffe, 1963), and instruments on instrumental activities of daily living (IADL) (McDowell & Newell, 1996). We also reviewed items on cognitive function from the Short Portable Mental Status Questionnaire (Pfeiffer, 1975). As a result of this process, a preliminary, closed-format, 27-item questionnaire was developed.
Reliability testing, final item derivation, and validation were conducted among three independent cohorts of patients. Eligible patients were new admissions to the Internal Medicine Service of a tertiary care academic center in New York City. Patients admitted to the HIV unit were excluded, because a multidisciplinary team including social workers evaluates all patients admitted to the HIV unit. The computerized database of patient admissions that is routinely used on our medicine services was used to generate a list of new admissions. Within 24 hours of admission, patients on that list were approached and invited to participate in the study. Information for patients who were cognitively impaired and who could not provide informed consent was obtained from a contact person or health care proxy written in the patient's chart (Ostbye, Tyas, McDowell, & Koval, 1997). All interviews were conducted by college-level trained research assistants. Overall, 93 percent of the patients who were approached agreed to participate in the study. Patients admitted from nursing homes were included in this study because they have greater rates of hospitalization, readmission, and utilization (Glass & Weiner, 1976; Murtaugh & Freiman, 1995). Prior screens have included admissions from nursing homes (Blaylock & Cason, 1992; Evans, Hendricks, Lawrence, & Bishop, 1988). The study protocol was approved by the Institutional Review Board at this institution.
To assess the reliability of the items on the questionnaire, 52 consecutive patients who were admitted to the medicine service were interviewed on two separate occasions. The degree to which repeated administrations of the questionnaire resulted in the same responses was determined using Cohen's weighted kappa (Dawson-Saunders & Trapp, 1994). Questionnaire items that had kappa of .50 or greater were considered to have an acceptable level of agreement. There are no standard criteria for selecting appropriate cut-offs for kappa values. Therefore, we used the guidelines proposed by Landis and Koch (1977b). According to their guidelines, a value of .40 to .60 is considered moderate (Kramer & Feinstein, 1981; Landis & Koch, 1977a, 1977b). Compared with other guidelines, these cut-off points are generous (Fleiss, 1981).
Derivation of Final Items
In the item derivation phase, the items remaining on the questionnaire after reliability testing were administered to a group of 299 patients admitted to the medicine service. Several variables have been identified as criteria for designating patients at risk of becoming disposition problems and who may therefore require a social work evaluation. However, no "gold standard" criteria exist. Therefore, the clinical judgments of social workers were used as the standard for determining whether a social work evaluation was needed. Feinstein (1996) proposed clinical judgment as an alternative to selecting items or determining importance when no external criteria exist. Although alternative measures such as length of stay (LOS) or discharge destination could have been used as outcome measures, the primary objective of the scale was to identify patients in need of early comprehensive discharge planning by a social worker. Therefore, the primary method of validation was the clinical judgment of social workers.
The three social workers who functioned as our panel of experts have master's degrees in social work from accredited schools of social work. One of the social workers is director of the Department of Social Work at our institution. The other two social workers hold administrative titles as program administrator and supervisor. Two of the social workers hold the certificates ACSW (Academy of Certified Social Workers) and QCSW (Qualified Clinical Social Worker). All three social workers have been involved in clinical social work and education for more than 10 years.
The social workers, who were blind to the patient's identity and to whether a patient had been seen by a social worker, were asked to review a standardized abstract of each patient's chart. Embedded in these abstracts were the candidate questionnaire items, which were not identified as potential variables for inclusion in the final questionnaire. These abstracts also contained the following information: demographics, reason for admission, comorbidity (Charlson, Pompei, Ales, & MacKenzie, 1987), recent hospitalizations, functional status, mental status, ADL and IADL impairments, financial status, and social support status.
Using this information, the social workers were asked the following question: Does this patient require a social work evaluation? They were asked to respond on a five-point Likert scale. The responses were as follows: definitely (56 percent), probably (14 percent), uncertain (3 percent), probably not (13 percent), and definitely not (14 percent). All three social workers reviewed each case. Differences of opinion most commonly occurred regarding whether a patient should be classified as "probably needed" or "definitely not needed" When this occurred, the social workers referred to a screening protocol that they routinely use to determine whether a new admission to the hospital should be evaluated. This protocol is a list of 16 items used on every admission. When a difference of opinion occurred, patients were classified as "probably needed" if they met two or more of the 16 criteria. If they met only one criterion, they were classified as "probably not needed." If an agreement could not be reached, the case was designated as "uncertain" which occurred in 3 percent of the 299 cases.
Based on the distribution of the social workers' responses, patients were grouped into two categories, those who definitely required an evaluation and those who did not. The category of uncertain was not included in these analysis; therefore, 56 percent were contrasted with 41 percent. Logistic regression was used to determine which items on the questionnaire were significantly associated with definitely needing a social work evaluation. Each item on the questionnaire was entered into a bivariate model, and those that were significant at the .01 level were retained for subsequent analysis. Among the remaining items, those that were highly correlated (r > .5) were eliminated or combined into one item. For example, four items that assessed ambulation were highly correlated and were therefore combined to form one hierarchical variable that was added to the logistic regression model. This was guided statistically by using cluster analysis, which allows for clustering of highly correlated variables. There is no standard hierarchical classification of walking mobility. The hierarchical classification of this item may bias the scores on the scale toward giving higher scores for those who are unable to walk across a room. However, the literature suggests that the inability to walk across a room or on a level surface indicates a greater degree of disability than the inability to climb stairs (Kirkwood, Culham, & Costigan, 1999; McDermott et al., 1999). The remaining items were entered as independent covariates into a multivariate logistic regression model with "definitely needing" a social work evaluation as the dependent variable. Items that remained significant in the multivariate model were chosen as the final items on the instrument. Although multivariate analysis can be performed, we performed multiple sequential bivariate analyses to further examine the clinical and statistical significance of each variable entered in the final model (Feinstein, 1996; Portney & Watkins, 2000).
Scoring of the Final Items on the Scale
To quantify the degree to which a social work evaluation was needed, a numerical scoring system was developed. In developing scales, logistic coefficients are often used to assign weights to items (Nunnally, 1978). However, multivariate odds ratios can also be used, because the exponentiated form of the regression coefficient is equal to the odds ratio. In addition, the use of odds ratios simplifies the scoring system by providing whole integers rather than decimals. In this study, each item on the scale was assigned a weight based on its multivariate odds ratio. An additional step was required in assigning a weight to item number one because it consisted of four components: the ability to walk a flight of stairs, the ability to walk a block, the ability to walk across a room, or the inability to do any of these tasks. Based on the results of cluster analysis and hierarchical categorical modeling, each component of item number one was assigned a numerical weight (1, 2, 3, or 4). The final score was obtained by adding the individual weights of all of the items. The final scores on this scale could range from 0 to 13, but in the derivation and validation population of this study, the scores ranged from 1 to 11. Higher scores on this scale indicated a greater need for a social work evaluation.
To determine whether an acceptable level of reliability would be retained when the items were combined to form a numerical score, we conducted an additional reliability analysis of the scale. The items that were tested in the reliability phase and that were retained on the final scale were retrieved from the reliability data. The scoring system was applied to these items. As previously stated, patients in the reliability phase completed the same questionnaires on two occasions; therefore each patient had two scores. The agreement between these scores was .7, demonstrating acceptable test-retest reliability when the items were combined.
Discriminative Ability of the Scale
Receiver operating characteristic (ROC) curves were used to identify optimal cut-off points. The overall discriminative ability of the model was determined by calculating the area under the curve (AUC). In addition, the coefficient of determination of the final scale was determined.
The validation of the SWAAT as an indicator of patients who have special discharge needs and who require a social work evaluation was based on the following hypotheses:
1. Patients who have complicated discharge planning needs have increased length of hospitalizations (Evans et al., 1988; Glass & Weiner, 1976; Huyse et al., 2001). Therefore, patients with higher scores on the SWAAT will have longer lengths of stay.
2. Patients with complicated discharge planning needs have a greater need for social services on discharge. Therefore, a greater proportion of patients with high scores on the SWAAT will require more services on discharge compared with patients with low or intermediate scores (Fairchild et al., 1998).
To validate the SWAAT we compared the proportion of patients requiring services on discharge in different score ranges. Finally, we compared the length of stay across each score range. All analyses were performed with SAS Version 6.12 software.
Fifty-two patients participated in the reliability-testing phase; their mean age was 57 (SD = 18 years), and 54 percent were female. Forty-six percent of patients were married, and 82 percent had completed high school. Eight percent of patients were uninsured, 37 percent had Medicare, 12 percent had Medicaid, and the remaining patients had private coverage. After reliability testing, 19 of the original 27 items remained on the questionnaire. These items had kappa values ranging from 0.50 to 1.0, which indicates a strong degree of agreement.
Derivation of Final Items
The characteristics of the 299 patients who participated in the item reduction phase are shown in Table 1. Table 2 shows the items that were found to be significant predictors of patients who definitely needed a social work evaluation in multivariate analysis. These items remained significant after adjusting for age, gender, ethnicity, and comorbidity. After reviewing the instrument, the study team decided to add an item that assessed whether the patient lived alone and was confused on admission. This item was included because baseline mental status and living situation are important determinants of length of stay and resource utilization. (Blaylock & Cason, 1992; Evans et al., 1988).
The six-item final scale (SWAAT) was validated among 200 patients admitted to the medicine service. Patients who participated in the validity phase were similar in demographic and clinical characteristics to patients in the reliability testing phase and the item derivation phase (Table 1). The range of scores on the SWAAT was from 1 to 11. The mean was 4.6 (SD = 2.1).
Based on ROC curves, a score of three or above was identified as having optimal discriminative ability. At this score the true positive rate was .80 and the false positive rate was .30. Using the method described by Cantor and Kattan (2000), the AUC was calculated as .75, which indicates good discriminative ability. We then compared this AUC with that of a prediction rule developed by Fairchild (Fairchild et al., 1998), which was specifically designed to predict the use of postdischarge services among medical inpatients. The prediction rule relies on age and scores on the physical and social function component of the Medical Outcomes Scale (Ware et al., 1996a, b). Using the Cantor and Kattan approach, in our study, the AUC for this prediction rule was .59. Although the AUC of the SWAAT was larger compared with that of the prediction rule proposed by Fairchild, there was no statistically significant difference between the two scales. The z statistic was .89 (Hanley & McNeil, 1983).
The ROC curve method provides an optimal diagnostic cutoff point, but it does not provide other stratifications (Feinstein, 1996). Therefore, a three-level stratification system was developed based on the distribution of length of stay and the proportion of patients requiring services in each score group. Increasing patterns of length of stay and need for services were observed among patients with scores ranging from 1 to 2, 3 to 6, and greater than 6. These categories were then applied in the validation phase. We used coefficient of determination ([r.sup.2]) to assess the ability of the model to identify patients who required social services. The coefficient of determination ([r.sup.2]) for this scoring system was .40, with an alpha level of .05. The association between score on the SWAAT and need for social work was significant (p < .001).
Twenty-four percent of patients with scores ranging from 1 to 2 were categorized as having low need for early social work evaluation. Fifty-six percent of patients had scores ranging from 3 to 6 and were categorized as having an intermediate need for early social work evaluation. The remaining 20 percent of patients had scores greater than 6 and were categorized as having a high need for early social work evaluation.
Of the patients enrolled in the validation phase, 141 (62 percent) were evaluated by a social worker. As the range of scores increased from low to intermediate to high, the proportion of patients who were evaluated by a social worker also increased from 33 percent to 67 percent to 87 percent, respectively. With an alpha level of .05, the association between score on the SWAAT and evaluation by a social worker was significant (p < .001). A total of 41 percent of patients required services at discharge. As the range of scores increased from low to intermediate to high, there was also an increase in the proportion of patients requiring services at discharge from 25 percent to 43 percent to 63 percent, respectively. With an alpha level of .05, the association between score on the SWAAT and discharge services was significant (p < .01). The most common services were visiting nurse services, home health attendant, and transfer to a skilled nursing home. When comparing length of stay across the three groups, as the scores increased from low to intermediate to high, the length of stay increased from six to 10 to 13 days respectively (see Figure 1). With an alpha level of .05, the association between score on the SWAAT and length of stay was significant (p < .01).
[FIGURE 1 OMITTED]
The SWAAT is a mechanism for identifying patients shortly after admission who have complicated social needs and who require a social worker evaluation. It focuses on six areas: difficulty with ambulation, living situation, current social services, patient's perception of the need for additional services, patient's perception of the need for help leaving the hospital, and whether the patient was noted to be confused at admission but lived alone. The reliability and validity of the items have been established. Items contained in this scale have been shown to be indicators of subsequent length of hospitalization and subsequent use of social services (Blaylock & Cason, 1992; Evans et al., 1988; Parfrey et al., 1994).
Because no gold standard exists for determining need for social work evaluations, we determined the specificity and sensitivity of the screen in explaining length of stay (LOS). When LOS was used as the gold standard, the sensitivity was 90 percent and specificity was 30 percent. Given the skewed nature of LOS in this population, the true state of disease, or in this case extended LOS, is not known. The high false-positive rate may be attributed to the skewed nature of LOS. Other patient factors, such as severity of illness, also may influence LOS. In the absence of a true gold standard test with which to compare the SWAAT, assessment of its diagnostic properties may be limited (Schulzer, 1994).
Based on the AUC approach, the discriminative ability of the SWAAT was determined to be .75, which indicates that the SWAAT can accurately distinguish between patients who need services and those who do not. Compared with the prediction rule developed by Fairchild, the SWAAT had a higher AUC, .75 versus .59. Although these differences are not statistically significant, the SWAAT has fewer items, which may make it easier to incorporate into the routine admission process with little burden on the patient and medical staff. Information can be gathered during the routine history taking or from proxy informants if the patient is cognitively impaired.
The SWAAT is comprehensive and includes items on the medical, functional, and social characteristics of patients. These items provide guidance in understanding and responding to the psychosocial needs of hospitalized patients. For example, assistance with ambulation can be provided, more services at home can be instituted, or placement in skilled nursing facilities can be coordinated for patients who are cognitively impaired but live alone. In designing the SWAAT the goal was to select items that had statistical and clinical relevance. The items that addressed confusion on admission and living situation were not statistically significant in this population; however, based on the literature, poor baseline mental status and living alone are significant predictors of high social risk and health care use (Berkman et al., 1980; Blaylock & Cason, 1992).
The SWAAT was designed to be administered early in the admission process. Ideally, it should be administered by those who have initial contact with patients so that a social work evaluation can be made quickly. The SWAAT can be administered by the nursing staff or physicians during the routine history assessment at the time of admission. No formal training is required to use the SWAAT. In our study, the scale was administered by college-level research assistants without formal training in social work or as health professionals.
The scoring system was designed so that a score can be generated at the bedside and then relayed to the appropriate social work service. Patients with the highest scores had the greatest need for a social work evaluation. As scores increased on the SWAAT from low to intermediate to high, there was a concomitant increase in the proportion of patients who were seen by a social worker, in the proportion of patients who received services at discharge, and in the proportion of patients who had increased length of stay. The scoring system also provides a method to stratify patients according to their needs and can help prioritize resources. If the caseload is low or the availability of social workers for discharge planning is high, the scores can be reprioritized and patients with lower scores can be evaluated sooner.
Although the SWAAT was designed for discharge planning, other applications can be tested in future research. The SWAAT provides a mechanism to standardize indications for social work referrals. In one survey more than one-half of the patients did not know why they were referred to a social worker or who had made the referral (Kadushin, 1996). Patients who meet the criteria for requiring discharge planning services can be told that they are being referred to the social work department based on their scores on the SWAAT. Scores can also be communicated during regular discharge planning rounds. In a survey of more than 300 hundred hospitals, cooperation from physicians was found to be an important predictor of perceived discharge planning effectiveness (Feather, 1993). Standardizing the concept of "need" can facilitate communication between the members of the discharge planning team.
Another potential use for the SWAAT is to describe the heavy caseload of hospital social workers in terms of degree of complexity in discharge planning. Coordinating services for patients with multiple needs in a short time period can be challenging. Discharge planning has been referred to as one of the most time consuming services provided by social workers (Coulton, Keller, & Boone, 1985). Higher scores suggest greater complexity in discharge planning. This may have implications for reimbursement of hospital social workers.
We have outlined the steps taken to develop and test the psychometric properties of the SWAAT. It is comprehensive, yet feasible and practical to use. The SWAAT serves as a case-finding tool that can be used to mobilize resources for patients exhibiting the greatest need. The SWAAT can be applied to a wide spectrum of medical inpatients, but testing of the SWAAT is needed in other patient populations, such as surgical or ambulatory care patients. The SWAAT provides a framework for designing tailored interventions that may reduce the amount of social stay days, reduce hospital expenditures, and improve patient outcomes.
Table 1: Characteristics of Patients in the Derivation and Validity Phase of Scale Development Derivation Validity Phase Phase (N = 299) (N = 200) Age (M [SD]) 63 (18) 70 (17) Female (%) 54.7 61.0 African American (%) 16.1 17.2 Latino American (%) 12.6 13.0 Did not complete high school (%) 24.1 25.5 Not currently married (%) 56.0 58.0 Unemployed (%) 9.1 4.5 Uninsured (%) 7.7 6.5 Medicare (%) 39.6 49.5 Medicaid (%) 14.4 11.5 Hospitalized < 30 days before admission (%) 22.4 28.5 Admitted from nursing home (%) 3.0 4.5 Hungry and unable to pay for food (%) 10.6 3.5 Comorbidity score > 3 (%) 50.7 46.5 Confused on admission, lives alone 3.7 3.0 Fair or poor self-perceived health (%) 70.0 70.0 Table 2: Multivariate Predictors of the Definite Need for Social Work Evaluation, According to the Blinded Social Work Assessment Odds Ratio Final Scale Items N (95% CI) Ft Weights 1. Able to walk: One flight of stairs 82 9.8 (5.0-19.2) .0001 1 Able to walk one block 36 2.9 (1.3-6.5) .0001 2 Able to walk across a room 68 3.4 (1.7-6.6) .0095 3 Not able to do any of the above 11 1.6 (1.3-2.1) .0003 4 2. Will not be able to return to the current place of residence 27 3.3 (1.1-10.1) .037 1 3. Has a home health aide, visiting nurse, or is in nursing home 72 41.4 (12.1-141.6) .0001 2 4. Needs additional help at home 54 4.7 (2.1-11.0) .0003 2 5. Needs help leaving the hospital and/or with medical visits 58 5.3 (2.4-11.8) .0001 2 6. Confused on admission and lives alone (a) 11 2.3 (0.2-24.9) .48 2 * Item 6 was added after the derivation phase based on clinical consensus.
Becker, N. E., & Becker, F. W. (1986). Early identification of high social risk. Health & Social Work, 11, 26-35.
Berkman, B., Rehr, H., & Rosenberg, G. (1980). A social work department develops and tests a screening mechanism to identify high social risk situations. Social Work in Health Care, 5, 373-385.
Blaylock, A., & Cason, C. L. (1992). Discharge planning predicting patients' needs. Journal of Gerontological Nursing, 18(7), 5-10.
Cantor, S. B., & Kattan, M. W. (2000). Determining the area under the ROC curve for a binary diagnostic test. Medical Decision Making, 20, 468-470.
Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987).A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases, 40, 373-383.
Coulton, C. J. (1988). Evaluating screening and early intervention: A puzzle with many pieces. Social Work in Health Care, 13, 65-72.
Coulton, C. J., Keller, S. M., & Boone, C. R. (1985). Predicting social workers' expenditure of time with hospital patients. Health & Social Work, 10, 35-44.
Creditor, M. C. (1993). Hazards of hospitalization of the elderly. Annals of Internal Medicine, 118, 219-223.
Dawson-Saunders, B., & Trapp, R. (1994). Basic and clinical biostatistics (2nd ed.). Norwalk, CT: Appleton and Lange.
Dobrof, J. (1991). DRGs and the social worker's role in discharge planning. Social Work in Health Care, 16, 37-54.
Evans, R. L., Hendricks, R. D., Lawrence, K. V., & Bishop, D. S. (1988). Identifying factors associated with health care use: A hospital-based risk screening index. Social Science and Medicine, 27, 947-954.
Fairchild, D. G., Hickey, M. L., Cook, E. F., McCarthy, R. M., Rossi, L. P., Timmons, S. C., Mangione, C. M., Lee, T. H. (1998). A prediction rule for the use of postdischarge medical services. Journal of General Internal Medicine, 13, 98-105.
Feather, J. (1993). Factors in perceived hospital discharge planning effectiveness. Social Work in Health Care, 19, 1-14.
Feinstein, A. R. (1996). Multivariable analyses: An introduction. New Haven, CT: Yale University Press.
Fillit, H., Howe, J. L., Fulop, G., Sachs, C., Sell, L., Siegel, P., Miller, M., & Butler, R.N. (1992). Studies of hospital social stays in the frail elderly and their relationship to the intensity of social work intervention. Social Work in Health Care, 18, 1-22.
Fleiss, J. (1981). Statistical methods for rates and proportions. New York: John Wiley & Sons.
Glass, R. I., & Weiner, M. S. (1976). Seeking a social disposition for the medical patient: CAAST, a simple and objective clinical index. Medical Care, 14, 637-641.
Hanley, J. A., & McNeil, B. J. (1983). A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology, 148, 839-843.
Holliman, D. C., Dziegielewski, S. F., & Datta, P. (2001). Discharge planning and social work practice. Social Work in Health Care, 32, 1-19.
Huyse, F. J., de Jonge, P., Slaets, J.P.J., Herzog, T., Lobo, A., Lyons, J. S., Opmeer, B. C., Stein, B., Arolt, V., Balogh, N., B., Cardoso, G., Fink, P., Rigatelli, M. (2001). COMPRI--An instrument to detect patients with complex care needs: Results from a European study. Psychosomatics, 42, 222-228.
Kadushin, G. (1996). Elderly hospitalized patients' perceptions of the interaction with the social worker during discharge planning. Social Work in Health Care, 23(1), 1-21.
Katz, S., Ford, A. B., Moskowitz, R.W., Jackson, B. A., & Jaffe, M. W. (1963). Studies of illness in the aged. The index of ADL: A standardized measure of biological and psychosocial function. JAMA, 185, 914-919.
Kirkwood, R. N., Culham, E. G., & Costigan, P. (1999). Hip moments during level walking, stair climbing, and exercise in individuals aged 55 years or older. Physical Therapy, 79, 360-370.
Kramer, M. S., & Feinstein, A. R. (1981). Clinical biostatistics. LIV. The biostatistics of concordance. Clinical Pharmacology and Therapeutics, 29, 111-123.
Landis, J. R., & Koch, G. G. (1977a). An application of hierarchical kappa-type statistics in the assessment of majority agreement among multiple observers. Biometrics, 33, 363-374.
Landis, J. R., & Koch, G. G. (1977b). The measurement of observer agreement for categorical data. Biometrics, 33, 159-174.
McDermott, M. M., Mehta, S., Liu, K., Guralnik, J. M., Martin, G. J., Criqui, M. H., & Greenland, P. (1999). Leg symptoms, the ankle-brachial index, and walking ability in patients with peripheral arterial disease. Journal of General Internal Medicine, 14, 173-181.
McDowell, I., & Newell, C. (1996). Measuring health. New York: Oxford University Press.
Murtaugh, C. M., & Freiman, M. P. (1995). Nursing home residents at risk of hospitalization and the characteristics of their hospital stays. Gerontologist, 35, 35-43.
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill.
Ostbye, T., Tyas, S., McDowell, I., & Koval, J. (1997). Reported activities of daily living: Agreement between elderly subjects with and without dementia and their caregivers. Age and Ageing, 26, 99-106.
Parfrey, P. S., Gardner, E., Vavasour, H., Harnett, J. D., McManamon, C., McDonald, J., & Dawe, J. (1994). The feasibility and efficacy of early discharge planning initiated by the admitting department in two acute care hospitals. Clinical and Investigative Medicine, 17, 88-96.
Pfeiffer, E. (1975). A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. Journal of the American Geriatrics Society, 23, 433-441.
Portney, L. G., & Watkins, M. P. (2000). Foundations of clinical research applications to practice (2nd ed.). Upper Saddle River, NJ: Prentice Hall Health.
Schulzer, M. (1994). Diagnostic tests: A statistical review. Muscle & Nerve, 17, 815-819.
Selker, H. P., Beshansky, J. R., Pauker, S. G., & Kassirer, J. P. (1989). The epidemiology of delays in a teaching hospital. The development and use of a tool that detects unnecessary hospital days. Medical Care, 27, 112-129.
Tennier, L. D. (1997). Discharge planning: An examination of the perceptions and recommendations for improved discharge planning at the Montreal General Hospital. Social Work in Health Care, 26, 4160.
Wachtel, T. J., Fulton, J. P., & Goldfarg, J. (1987). Early prediction of discharge disposition after hospitalization. Gerontologist, 27, 98-103.
Ware, J., Kosinski, M., & Keller, S. D. (1996a). A 12-item short-form health survey: Construction of scales and preliminary tests of reliability and validity. Medical Care, 34, 220-233.
Ware, J, Kosinski, M., & Keller, S. D. (1996b). How to score the SF-12 Physical and Mental Health Summary Scales. Boston: New England Medical Center, Health Institute.
Original manuscript received June 4, 2001 Final revision received September 19, 2002 Accepted April 4, 2003
Carla Boutin-Foster, MD, MS, is an internist, Division of General Internal Medicine, Weill Medical College of Cornell University, 525 East 68th Street, Box 46, New York, NY 10021; e-mail: firstname.lastname@example.org. Sona Euster, MSSA, CSW, is director, Yvette Rolon, MSSW, ACSW, CSW, is program adminstrator, and Athena Motal, CSW, is social work manager, Department of Social Work, New York Presbyterian Hospital-Weill Cornell Center, New York. Rhonda BeLue, PhD, is a consulting statistician, Weill Medical College of Cornell University, New York. Robin Kline, MS, is administrative director, Kreitchman PET Center, College of Physicians and Surgeons, Columbia University, New York. Mary E. Charlson, MD, is chief, Division of General Internal Medicine, Weill Medical College of Cornell University, New York. Funding was provided by a grant from the New York Community Trust and the New York Presbyterian Hospital. An earlier version of this article was presented at the 3rd annual meeting of the Society for Clinical Epidemiology and Health Care Research, May 2000, Boston.
|Printer friendly Cite/link Email Feedback|
|Author:||Boutin-Foster, Carla; Euster, Sona; Rolon, Yvette; Motal, Athena; BeLue, Rhonda; Kline, Robin; Charl|
|Publication:||Health and Social Work|
|Date:||May 1, 2005|
|Previous Article:||Communication between older people and their health care agents: results of an intervention.|
|Next Article:||Ethical considerations in prenatal sex selection.|
|Use of a quality of life instrument to improve assessment of brain tumor patients in an outpatient setting.|
|Education: a complex and empowering social work intervention at the end of life.|