• 沒有找到結果。

Individual and Hospital Factors Associated with Hospitalization for Chronic Medical Conditions

N/A
N/A
Protected

Academic year: 2021

Share "Individual and Hospital Factors Associated with Hospitalization for Chronic Medical Conditions"

Copied!
13
0
0

加載中.... (立即查看全文)

全文

(1)9. Individual and Hospital Factors Associated With Hospitalization for Chronic Medical Conditions Cheng-Chieh Lin1, Leiyu Shi2, Tsai-Chung Li1,3 1. Department of Family Medicine, Center of Community Medicine, China Medical College. Hospital; 3Institute of Chinese Medicine, China Medical College, Taichung, Taiwan, R.O.C. 2. Primary Care Policy Center, Department of Health Policy and Management, School of Hygiene and Public Health, Johns Hopkins University, Baltimore, Maryland, U.S.A.. This study identified individual and hospital characteristics significantly associated with U.S. hospital admissions for chronic medical conditions (i.e., asthma, hypertension, congestive heart failure, chronic obstructive pulmonary disease, and diabetes) and assessed whether the results based on hospital admissions for chronic medical conditions were consistent with analysis based on hospital admissions for ambulatory care sensitive conditions (ACSC). Data for this study were from the 1994 U.S. National Hospital Discharge Survey. Bivariate statistical comparisons were performed to test the differences between hospitalized individuals with chronic medical conditions and those without. Logistic regression followed to determine the significance of independent variables in relation to hospitalizations for chronic medical conditions. The logistic regression results of using hospital admissions for chronic medical conditions as the dependent variable and using hospital admissions for ACSC as the dependent variable was compared to examine the level of congruence for the two models using different dependent measures. Individual and hospital characteristics significantly associated with hospital admissions for chronic medical conditions (objective 1) included age, gender, race, marital status (individual predisposing factors), principal and secondary sources of payment (individual enabling factors), length of stay (individual need factor), and number of hospital beds and geographic region (system factors). The results of analyses based on hospital admissions for chronic medical conditions were consistent with analysis based on hospital admissions for ACSC (objective 2). Hospital admissions for chronic medical conditions can serve as an efficient way of identifying subpopulations facing access barriers. ( Mid Taiwan J Med 1999 4 9-21 ). Key words chronic medical conditions, factor, hospitalization. INTRODUCTION. colleagues defined ACSC as diagnoses for. The analysis of variations in hospitalization. which timely and effective outpatient or. rates for ambulatory care sensitive conditions. ambulatory care can help reduce the risks of. (ACSC) has emerged as an alternative measure. hospitalization by either preventing the onset. of health care access [1-6]. Billings and his. of an illness or condition, or managing a. Received August 13, 1998.. chronic illness or condition [1]. Heretofore,. Revised October 1, 1998.. Accepted November 2, 1998.. researchers have relied on state-wide hospital. Address reprint requests to Cheng-Chieh Lin, Department of. discharge datasets for the analysis. For. Family Medicine, Center of Community Medicine, China Medical College Hospital, No 2, Yuh-Der Road, Taichung, Taiwan, R.O.C.. example, Weissman et al. found that the uninsured. and. Medicaid. patients. in.

(2) Individual and Hospital Factors for Chronic Medical Conditions. 10. Massachusetts and Maryland were more likely. seeking, and physician practice style. They. to be admitted to a hospital for chronic. concluded that low-income communities. medical conditions (CMC) than privately. where people perceived poor access to. insured patients [3]. Billings and his associates. medical care had higher rates of preventable. studied patterns of hospital use in New York. hospitalization and hospitalization for CMC.. City and found that the hospitalization rates. The first objective of this study was to. for ACSC were higher in low-income areas. identify individual and hospital characteristics. than in higher income areas where appro-. significantly. priate outpatient or ambulatory care was. admissions for CMC in the United States (U.S.).. more readily available [1,4]. The Codman. Given the linkage between access to care and. Research Group conducted a 15-state com-. hospital admis-sions for CMC, the results of the. parative study and noted that per capita. study would produce a profile of individuals. admission rates for ACSC were directly related. experiencing access barriers to primary and. to poverty rates for all states with a significant. ambulatory care, thus assisting policy makers. urban population [5].. in developing programs that aim at improving. associated. with. hospital. The current study aimed at expanding the. access to care for these individuals. The. analysis by using a national hospital discharge. second objective was to assess whether the. dataset and focusing on a set of CMC. results based on hospital admissions for CMC. including asthma, hypertension, congestive. were consistent with results based on hospital. heart failure, chronic obstructive pulmonary. admissions for ACSC as identified by Billing. disease, and diabetes. Analysis based on. and associates. Specifically, the results of the. national data will be more generalizable than. analyses using hospital admissions for CMC. state level studies. The choice of a subset of. and for ACSC were compared. The congruence. ACSC identified by Billing and his colleagues. of the results would indicate that using. was due to two factors. First, CMC such as. hospital admissions for CMC could serve as an. those identified can often be treated more. efficient way of identifying subpopulations. timely, effectively, and efficiently in an. facing access barriers.. ambulatory setting rather than inpatient setting. Therefore, hospitalization for these conditions generally indicates lack of access to ambulatory care. Second, the validity of using variations in CMC hospitalization rates as measure of access to care has been established by previous researchers. Bindman et al. assessed the validity of using CMC as measure of health care access using California Hospital Discharge Data and a random digit telephone survey [6]. They compared respondents’ reports of access to medical care in an area with the area’s hospital admission rate for five CMC, namely, asthma, hypertension, congestive heart failure, chronic obstructive pulmonary disease, and diabetes. The results indicated that access to care was inversely associated with hospitalization rates for the five CMC after controlling for the prevalence of the conditions, health care. MATERIALS AND METHODS Data. The 1994 National Hospital Discharge Survey (NHDS) was used as the primary data source for conducting the analysis. The NHDS is conducted annually by the National Center for Health Statistics (NCHS) and is a principal source of information on inpatient hospital utilization in the U.S. This survey collects medical and demographic information from a sample of discharge records selected from non-institutional hospitals, exclusive of Federal, military, and Veterans Administration hospitals, located in the 50 states and the District of Columbia. Only short-stay hospitals (hospitals with an average length of stay for all patients of less than 30 days) or those that specialize in general care (medical or surgical) or pediatric general care are included in the.

(3) Cheng-Chieh Lin, et al.. 11. survey. The hospitals surveyed have six or. insignificant association between income level. more beds staffed for patients’ use.. and hospital admissions for ACSC among. Data collection included both manual and. elderly patients [1,4]. Obstetrical patients with. automated procedures. The manual system of. normal delivery were excluded because. sample selection and data abstraction was. normal delivery was not considered an illness.. used for approximately 62% of the responding. Variables. hospitals. The automated procedure involved. For the purpose of this study, we selected. the purchase of computerized data tapes from. variables. abstracting service organizations, state data. characteristics associated with hospitalization.. systems, or from the hospitals themselves. This. The purpose was to find out which. method was used for approximately 38% of. characteristics were significantly related to. the respondent hospitals. The system used for. variations in hospitalizations for CMC and if. coding the diagnoses and procedures on the. these characteristics were also significantly. medical abstract forms as well as on the. associated with hospitalization for ACSC.. of. individual. and. hospital. commercial abstracting services data tapes was. With respect to the first research objective,. the International Classification of Diseases,. the dependent variable examined was. Ninth Revision, Clinical Modification (ICD-9-. discharge diagnoses grouped as CMC or non-. CM) [7].. CMC. The five CMC selected were based on. Detailed descriptions of the sampling. the work by Bindman and his colleagues [6].. design and data collection procedures have. The conditions were asthma, diabetes (A, B,. been published elsewhere [8,9]. The 1994. C), chronic obstructive pulmonary disease,. sample consisted of 525 hospitals. Of these, 13. congestive heart failure, and hypertension. To. were found to be ineligible because they went. classify. out of business or otherwise failed to meet the. conditions, both primary and secondary. criteria for the NHDS. Of the 512 eligible. diagnoses based on the ICD-9-CM code were. hospitals, 478 responded to the survey with a. used. The code specifications for the chronic. response rate of 93%.. medical conditions studied are listed in Table. Consistent with previous related research. the. hospitalization. for. these. 1.. [1,4,6], our study limited the analysis to adults. With respect to the second research. aged 18-64 who were formally admitted to the. objective, the second dependent variable. inpatient services of a short-stay hospital. examined in this study was discharge. surveyed in 1994 for observation, care,. diagnoses grouped as ACSC or non-ACSC. Only. diagnosis, or treatment. The pediatric (age 0 to. the primary diagnosis was used for classi-. 17), elderly (age 65 years or more), and. fication of the hospitalization for medical. obstetrical patients with completely normal. conditions. The selection of diagnoses for. deliveries were excluded from the analysis.. ACSC conditions was based on the listing of. The pediatric and elderly patients were. ICD-9-CM codes for ACSC developed by. excluded because it is likely that the ACSC. Billings and his colleagues [1,4]. A medical. applicable to children are different from those. advisory panel of internists and pediatricians,. applicable to adults. Analyses on the pediatric. including experts on access barriers, developed. patients are published elsewhere. The elderly. a diagnostic framework for analyzing hospital. population was excluded due to Medicare. use patterns [4]. Using the Delphi approach,. coverage which pays for a significant amount. they grouped hospital admissions into ACSC. of outpatient medical care and provides. and marker conditions, diagnoses for which. adequate reimbursement level for most. the provision of timely and effective. physicians to accept Medicare patients.. outpatient care is likely to have little impact. Previous. on the need for hospital admission.. research. on. ACSC. indicated.

(4) Individual and Hospital Factors for Chronic Medical Conditions. 12. Table 1 . Code specifications for the chronic medical conditions Chronic Conditions. ICD-9-CM Code. Asthma Diabetes A, B, C. 493 250.1, 250.2, 250.3, 250.8, 250.9, 250.0 491, 492, 494, 496, 466.0. Chronic obstructive pulmonary disease Congestive heart failure Hypertension. Exclusions. 428, 402.01, 402.11, 402.91, 518.4 401.0, 401.9, 402.00, 402.10, 402.90. Congestive heart failure cases with surgical procedures: 36.01, 36.02, 36.05, 36.1, 37.5, or 37.7 Hypertension cases with surgical procedures: 36.01, 36.02, 36.05, 36.1, 37.5 or 37.7. The independent variables used in our. variables was either very small or non-. study were demographics, geographic regions,. significant when need variables were taken. hospital and hospitalization characteristics, and. into account [11,12]. Excluding these two. sources of payment for inpatient care.. variables did not have much effect on the. Specifically, individual demographics included. results. Insurance status or sources of payment. age, sex, race and marital status. Geographic. was used as a measure of enabling factor.. region included Northeast, Midwest, South and. Specifically, sources of payment included both. West. The hospital and hospitalization. principal and secondary sources of payment. characteristics included number of beds in a. for inpatient care. Length of hospital stay was. hospital, hospital ownership, and days of. used as a proxy for need. The system. inpatient care (length of stay). Sources of. characteristics were limited to hospital related. payment for inpatient care were measured. measures available in the dataset, including. using the principal expected source of. the number of beds in a hospital, hospital. payment. A new independent variable, the. ownership, and geographic region such as the. additional source of payment, was created. Northeast, Midwest, South, and West. Table 2. from variables of principal expected source of. provides the operational definitions of the. payment and secondary expected source of. measures used in the analysis.. payment which have the same categories.. Analysis. Table 2 provides the operational definitions of the measures used in the analysis.. The analytical strategy in relation to the first research objective was to examine patient. The independent variables used in this. and hospital characteristics associated with. study were based on the framework for access. hospital admissions for CMC. First, descriptive. to care developed by Aday and Andersen [10].. statistics (i.e., means, standard deviations and. The framework conceptualizes access to care. distributions) were generated to provide a. as determined by both the characteristics of. profile of the general characteristics of the. the individuals at risk and the delivery system.. measures used in the study (Table 2). Next,. Individual characteristics include predisposing,. bivariate. enabling,. System. performed to test the differences between. characteristics include resources, volume,. individuals admitted due to CMC and those. distribution, organization, entry, and structure. due to other conditions. Chi-square tests were. measures. For the purpose of this study, the. used for categorical independent variables and. predisposing individual characteristics used. t-tests for continuous measures. Logistic. were age, sex, race, and marital status.. regression followed to determine the signi-. Information of occupation and education was. ficance of independent variables in relation to. not available in the dataset. The variation of. dependent variables. access explained by these two predisposing. five CMC.. and. need. factors.. statistical. comparisons. were. hospitalizations for the.

(5) Cheng-Chieh Lin, et al.. 13. Table 2. Definitions, means, standard deviations, and distributions of variables used in the analysis (N=125,621) Variables. Description. Distribution*. Age. The age of the patient on the birthday prior to admission to the hospital inpatient service Sex 1 = male 2 = female Race 1 = White 2 = Black 3 = American Indian/Eskimo 4 = Asian/Pacific Islander 5 = Other 9 = Not stated Marital status 0 = Married 1 = Other, including single, widowed, divorced, separated, unknown and, not stated Geographic region 1 = Northeast, includes Maine, New Hampshire, Vermont Massachusetts, Rhode Island, Connecticut, New York, New Jersey, and Pennsylvania 2 = Midwest, includes Michigan, Ohio, Illinois, Indiana, Wisconsin, Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas 3 = South, includes Delaware, Maryland, District of Columbia, Virginia, West Virginia, North Carolina, South Carolina, Georgia, Florida, Kentucky, Tennessee, Alabama, Mississippi, Arkansas, Louisiana, Oklahoma, and Texas 4 = West, includes Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, Nevada, Washington, Oregon, California, Hawaii, and Alaska Number of beds 1 = 6-99 2 = 100 - 199 3 = 200 - 299 4 = 300 - 499 5 = 500 and over Hospital ownership The type of organization that controls and operates the hospital 1 = Proprietary: Hospitals operated by individuals, partnerships, or corporations for profit 2 = Government: Hospitals operated by State and local government 3 = Nonprofit: Hospitals operated by a Church or another not for profit organization Principal expected 0 = No charge source of payment 1 = Workmen compensation 2 = Medicare 3 = Medicaid 4 = Other government payments, including Title V 5 = Blue Cross 6 = Other private/commercial insurance 7 = Self-Pay 8 = Other 9 = Not stated Additional source 1= With expected secondary source of payment of payment 0 = Without expected secondary source of payment Length of stay The total number of patient days accumulated at time of discharge by patients discharged from shortstay hospitals during a year. 46794 78827 71646 19153 834 2431 6638 24919 28685 96936. (37.3%) (62.7%) (57.0%) (15.2%) (0.7%) (1.9%) (5.3%) (19.8%) (22.8%) (77.2%). 33460. (26.6%). 34888. (27.8%). 37832. (30.1%). 19441. (15.5%). 13934 21959 28520 39028 22180. (11.1%) (17.5%) (22.7%) (31.1%) (17.7%). 11903. (9.5%). 14174. (11.3%). 99544. (79.2%). 928 2217 11346 24040 2589 18745 47274 8854 6826 2802 10107 115514. (0.7%) (1.8%) (9.0%) (19.1%) (2.1%) (14.9%) (37.6%) (7.0%) (5.4%) (2.2%) (8.0%) (92.0%). Mean. Max.. (SD). (Min.). 40.27 (13.36). 64.00 (18.00). 5.08 (7.99). 383.00 (1.00).

(6) Individual and Hospital Factors for Chronic Medical Conditions. 14 Table 2. Continued Variables. Description. Chronic conditions. 1 = With chronic conditions based on primary and 14664 secondary diagnoses 0 = Without chronic conditions based on primary 110957 and secondary diagnoses With asthma based on primary and secondary 2011 diagnoses With CHF based on primary and secondary 2584 diagnoses With COPD based on primary and secondary 2779 diagnoses With diabetes A, B, or C based on primary and 3698 secondary diagnoses. With hypertension based on primary and secondary 3592 diagnoses.. Asthma CHF COPD Diabetes A, B, C Hypertension. Distribution*. Mean. Max.. (SD). (Min.). (11.7%) (88.3%) (13.7%) (17.6%) (19.0%) (25.2%) (24.5%). *For categorical variables, the frequencies and percentage distributions are provided.. The chronic conditions. include asthma, chronic heart failure, chronic obstructive pulmonary disease, diabetes A, B, C, and, hypertension. CHF = congestive heart failure; COPD = chronic obstructive pulmonary disease.. The categorical variables were coded as. hospital characteristics were entered in both. sets of dummy variables in the logistic. models.. regression. Race was classified as White. coefficients and the odds ratios (OR) of the. (default category), Black, Asian/Pacific. significant. Islander, and other (including American. compared in the two models to examine the. Indian/Eskimo, other race, and those “not. level of congruence between the two models. stated”). Marital status was classified as Married. using different dependent variables.. The. standardized. parameter. regression. estimates. were. (default category) and Non-married (including single,. widowed,. divorced,. RESULTS. separated,. unknown, and those “not stated”). Principal. Table 2 provides the definitions and. source of payment was classified as No. descriptive statistics of the variables used in. charge,. Workmen. our study. In terms of individual charac-. compensation (including other government. teristics, the mean age of the patients was 40. payment), Private insurance (including Blue. years. Almost two-thirds (62.7%) of the. Cross. Medicare,. and. other. Medicaid,. private/commercial. patients were female and 22.8% were married.. insurance) (default category), Self-pay, and. White patients accounted for 57%, followed by. Other (including those “not stated”). Additional. Blacks (15.2%), Asian/Pacific Islanders (1.9%),. source of payment were grouped as those. American Indians/Eskimos (0.7%), other. who had secondary expected source of. (5.3%), and not stated (19.8%). The average. payment and those who did not (default. length of hospital stay was 5.1 days. Most. category).. patients (52.5%) had private insurance and 32%. The analytical strategy in relation to the. of patients had public insurance. Eight percent. second research objective was to compare the. of the patients also had secondary source of. logistic regression results of using hospital. payment.. admissions for CMC as the dependent variable. In terms of hospital characteristics, most. and using hospital admissions for ACSC as the. hospitals (79.2%) were nonprofit, followed by. dependent variable. The same patient and. government (11.3%) and proprietary (9.5%)..

(7) Cheng-Chieh Lin, et al.. 15. There were more patients from the Southern. Table 4 shows the results of two logistic. region (30.1%) than from other regions. The. regression models associating patients’. bed distribution of the hospitals was: 11.1% of. individual (i.e., age, sex, race, marital status,. the hospitals had 6-99 beds, 17.5% had 100-199. principal and secondary expected sources of. beds, 22.7% had 200-299 beds, 31.1% had 300-. payment, length of stay) and hospital. 499 beds, and 17.7% had 500 or more beds.. characteristics (i.e., ownership, geographic. Among the 125,621 adult discharged. region, number of beds) with CMC and ACSC. patients included in the analysis, 11.7% were. respectively. The OR of the significant. diagnosed with at least one of the five. parameter estimates are provided.. selected CMC (5.3% based on principal. The significant individual factors associated. diagnosis and 6.4% based on secondary. with having a CMC included age, sex, race, and. diagnosis). Among those diagnosed with. marital status. Specifically, controlling for other. chronic conditions, 13.7% had asthma (10.5%. individual. based on principal diagnosis and 3.2% based. individuals were more likely than younger. on secondary diagnosis), 17.6% had congestive. ones to have a CMC [OR=1.07; 95% confidence. heart failure (10.8% based on principal. interval (CI)=1.05, 1.09]. The odds of having a. diagnosis and 6.8% based on secondary. CMC for men were 1.10 times greater than the. diagnosis), 19% had chronic obstructive. odds for women (OR=1.10; 95% CI=1.06, 1.14).. pulmonary disease (7.5% based on principal. The odds of having a CMC for blacks were. diagnosis and 11.5% based on secondary. 1.63 times greater than the odds for whites. diagnosis), 25.2% had diabetes (11.0% based on. (95% CI=1.55, 1.71). Asians were less likely than. principal diagnosis and 14.2% based on. whites to have a CMC (OR=0.70; 95% CI=0.59,. secondary. had. 0.83). In comparison with those who were. hypertension (5.6% based on principal. married, non-married patients were more. diagnosis and 18.9% based on secondary. likely to have a CMC (OR=1.09; 95% CI=1.04,. diagnosis).. 1.15). Individuals’ insurance status was also. diagnosis),. and. 24.5%. and. hospital. factors,. older. Table 3 lists the five CMC and the individual. significantly and independently associated. and hospital characteristics associated with. with admission with CMC. Specifically,. them. Except for the patients with asthma, age. compared with those with private insurance,. and sex were significant variables related to. those without insurance (i.e., self-pay) had 1.45. admission. chronic. times the odds in favor of having a CMC. conditions; older people were more likely to. (OR=1.45; 95% CI=1.48, 1.70). Individuals with. be admitted than younger ones and males. Medicare or Medicaid were also significantly. were more likely to be admitted than females.. more likely to be admitted with a CMC. Those. Except for the patients with pulmonary. without secondary sources of payment were. disease, blacks were more likely to be. more likely to have a CMC than those with. admitted with preventable chronic conditions. secondary sources of payment (OR=1.24; 95%. than other races. Non-married individuals. CI=1.15, 1.32). Individuals with shorter lengths. were more likely to be hospitalized for asthma. of stay were more likely to be admitted with. and. a CMC (OR=0.99; 95% CI=0.98, 0.99).. with. pulmonary. preventable. disease. than. married. individuals who were more likely to be. With respect to hospital characteristics, the. hospitalized for hypertension. Hospitals in the. geographic location, hospital ownership, and. west were the least likely to have patients. number of beds were all independently. admitted with preventable chronic conditions.. associated. Hospitals with 200-299 beds were more likely. admissions. Hospitals in the Northwest,. to have patients admitted with preventable. Midwest, and South were more likely to have. chronic conditions.. individuals admitted for CMC than those in. with. preventable. hospital.

(8) Individual and Hospital Factors for Chronic Medical Conditions. 16. Table 3. Bivariate statistics for chronic conditions by individual and hospital characteristics Asthma Yes Age*. Sex Male Female. Race White Black American Indian /Eskimo Asian/Pacific Islander Other. Marital status Married Non-married. Region Northeast Midwest South West Number of beds 6-99 100-199 200-299 300-499 500 and over Hospital ownership Proprietary Government Nonprofit. No. 40.84 39.01 (12.72) (13.08) p>0.05. Diabetes Yes. No. COPD Yes. CHF No. Yes. Hypertension No. 47.50 39.00 53.12 39.00 54.39 39.01 (12.31) (13.08) (9.98) (13.08) (8.75) (13.08) p<0.01 p<0.01 p<0.01. 1.41 98.59 1.99 98.01 p>0.05. 4.25 2.64. 95.75 97.36 p<0.01. 1.65 2.63 2.52. 98.35 97.37 97.48. 2.99 4.78 1.02. 0.87. 99.13. Yes. No. 51.46 39.01 (9.26) (13.08) p<0.01. 3.29 96.71 1.96 98.04 p<0.01. 3.31 1.68. 96.69 98.32 p<0.01. 97.01 95.22 98.98. 2.92 1.63 1.02. 97.08 98.37 98.98. 2.07 3.61 1.02. 97.93 96.39 98.98. 2.98 4.40 2.03. 97.02 95.60 97.97. 1.56. 98.44. 0.78. 99.22. 1.39. 98.61. 2.15. 97.85. 1.98 98.02 p<0.01. 2.99. 97.01 p<0.01. 1.12. 98.88 p<0.01. 1.80. 98.20 p<0.01. 2.57. 97.43 p<0.01. 1.39 98.61 1.89 98.11 p<0.01. 3.07 96.93 3.27 96.73 p>0.10. 2.27 97.73 2.49 97.51 p<0.05. 2.22 2.29. 97.78 97.71 p>0.10. 3.75 2.95. 96.25 97.05 p<0.01. 2.36 97.64 1.91 98.60 1.40 98.60 1.27 98.73 p<0.01. 2.77 97.23 3.69 96.31 3.55 96.45 2.57 97.43 p<0.01. 2.06 97.94 3.20 96.80 2.44 97.56 1.78 98.22 p<0.01. 1.98 2.56 2.55 1.76. 98.02 97.44 97.45 98.24 p<0.01. 2.97 97.03 3.19 96.81 3.64 96.36 2.36 97.64 p>0.10. 1.72 98.28 1.56 98.44 2.03 97.97 1.84 98.16 1.63 98.37 p<0.01. 2.98 3.34 3.47 3.20 3.00. 97.02 96.66 96.53 96.80 97.00 p<0.01. 3.06 96.94 2.82 97.18 2.66 97.34 2.39 97.61 1.49 98.51 p<0.01. 2.10 2.01 2.58 2.31 2.20. 97.90 97.99 97.42 97.69 97.80 p<0.01. 3.14 2.97 3.35 3.07 3.14. 96.86 97.03 96.65 96.93 96.86 p>0.10. 1.23 98.77 1.72 98.28 1.85 98.15 p<0.01. 3.88 3.52 3.10. 96.12 96.48 96.90 p<0.01. 2.65 97.35 2.26 97.74 2.45 97.55 p>0.10. 3.04 96.96 2.07 97.93 2.21 97.79 p<0.01. 3.23 3.34 3.10. 96.77 96.66 96.90 p>0.10. 5.88 2.05. 94.12 97.95. 2.73 1.58. 97.27 98.42. 3.33 0.24. 96.67 99.76. 4.39 3.11. 95.61 96.89. 6.12 2.88 3.16. 93.88 97.12 96.84. 6.10 2.23 2.21. 93.90 97.77 97.79. 7.00 2.20 2.42. 93.00 97.80 97.58. 3.58 1.78 2.70. 96.42 98.22 97.30. 3.26 2.61. 96.74 97.39. 2.71 1.84. 97.29 98.16. 2.10 1.59. 97.90 98.41. 4.05 3.38. 95.95 96.62. 4.14 2.96. 95.86 97.04 p<0.01. 1.87 98.13 1.74 98.16 p<0.01. 3.13 3.12. 96.87 96.88 p<0.01. Source of payment No charge 1.26 98.74 Workmen 0.58 99.42 compensation Medicare 2.16 97.84 Medicaid 2.09 97.91 Other government 1.67 98.33 payment Blue Cross 1.82 98.18 Other private 1.52 98.48 insurance Self-pay 2.47 97.53 Other 1.63 98.37 p<0.01. 1.89 98.11 2.36 97.64 p<0.01. 4.38 95.62 2.42 97.58 p<0.01.

(9) Cheng-Chieh Lin, et al.. 17. Table 3. Continued Asthma Yes Length of stay* (days). Diabetes. No. 4.55 5.02 (4.82) (8.11) p<0.01. Yes. COPD. No. 5.35 5.02 (6.46) (8.11) p<0.01. Yes. CHF No. Hypertension. Yes. 6.31 5.02 (7.97) (8.11) p>0.10. No. 7.39 5.02 (8.94) (8.11) p<0.01. Yes. No. 4.26 5.02 (5.71) (8.11) p<0.01. * Means and standard deviation are provided. Each chronic condition is compared with control group (excluding any other four chronic conditions). COPD = chronic obstructive pulmonary disease; CHF = congestive heart failure.. Table 4.. Logistic regression results of individual and hospital characteristics associated with preventable. hospitalizations Diagnosed with. Diagnosed with. chronic medical conditions Independent variable. (SE). Age 0.07 (0.00) Sex (0-1) Male 0.10 (0.02) Race* Black 0.49 (0.03) Asian -0.35 (0.09) Other -0.04 (0.02) Marital status (0-1) Non-married 0.09 (0.02) Length of stay -0.01 (0.00) Geographic region Northwest 0.16 (0.03) Midwest 0.37 (0.03) South 0.27 (0.03) Principal source of payment No charge 0.47 (0.10) Workmen compensation -0.03 (0.05) Medicare 0.46 (0.03) Medicaid 0.36 (0.03) Self-pay 0.37 (0.04) Other 0.04 (0.03) No Secondary source 0.21 (0.03) of payment Hospital ownership Government 0.06 (0.04) Nonprofit -0.00 (0.03) Number of beds ** 6-99 0.22 (0.04) 100-199 0.24 (0.03) (0.03) 200-299 0.25 300-499 0.14 (0.03) Intercept -5.90 -2 LOG likelihood 90,538.30 Sample size 125,621 * Default category is White.. OR. ambulatory care sensitive conditions. (95% CI). (SE). OR (95% CI). 1.07. (1.05, 1.09). 0.04. (0.00). 1.04. (1.02, 1.06). 1.10. (1.06, 1.15). 0.10. (0.02). 1.10. (1.06, 1.15). 1.63 (1.54, 1.73) 0.49 0.70 (0.59, 0.84) -0.23 0.96 (0.92, 1.00) 0.01. (0.02) (0.08) (0.02). 1.63 (1.57, 1.70) 0.79 (0.68, 0.93) 1.01 (0.97, 1.05). 1.09 (1.05, 1.14) 0.08 0.99 (0.98, 0.995) 0.02. (0.02) (0.00). 1.08 (1.04, 1.13) 0.98 (0.97, 0.99). 1.18 1.45 1.31. (0.03) (0.03) (0.03). 1.28 1.35 1.27. (1.11, 1.25) 0.25 (1.37, 1.54) 0.30 (1.24, 1.39) 0.24. (1.21, 1.36) (1.27, 1.43) (1.20, 1.35). 1.60 (1.32, 1.95) 0.97 (0.88, 1.07) 1.58 (1.49, 1.68) 1.43 (1.35, 1.52) 1.45 (1.34, 1.57) 1.04 (0.98, 1.10) 1.24 (1.16, 1.31). 0.37 -0.22 0.45 0.33 0.46 0.06 0.12. (0.10) (0.06) (0.03) (0.03) (0.03) (0.04) (0.03). 1.44 (1.19, 1.76) 0.80 (0.71, 0.90) 1.56 (1.48, 1.66) 1.39 (1.31, 1.48) 1.59 (1.49, 1.68) 1.07 (0.98, 1.15) 1.12 (1.06, 1.20). 1.07 1.00. (0.98, 1.15) (0.94, 1.06). 0.24 0.04. (0.04) (0.03). 1.27 1.04. (1.18, 1.38) (0.98, 1.10). 1.25 1.27 1.28 1.16. (1.15, 1.35) (1.20, 1.35) (1.21, 137) (1.09, 1.22). 0.46 (0.03) 0.33 (0.03) 0.35 (0.03) 0.20 (0.03) -4.66 91,485.22 125,621. 1.58 1.40 1.41 1.22. (1.68, 1.49) (1.31, 1.48) (1.34, 1.51) (1.15, 1.30). Default category is West.. Default category is Private insurance.. is Proprietary. ** Default category is bed number of 500 and over.. p < 0.10;. p < 0.05;. p < 0.01,. Default category two-sided..

(10) Individual and Hospital Factors for Chronic Medical Conditions. 18. the West. Hospitals with the number of beds. were older, male, black, non-married, without. less than 500 were more likely to have. insurance or with Medicaid and Medicare, and. individuals admitted for CMC than hospitals. without expected secondary sources of. with 500 or more beds. For example, the odds. payment. These are population groups most. of admitting individuals with CMC for. likely to face access barriers to ambulatory. hospitals with 6-99 beds were 1.25 times. and primary care. Hospitals likely to have a. greater than for hospitals with 500 or more. higher rate of preventable hospitalizations. beds (95% CI=1.16, 1.34) after controlling for. were relatively smaller as measured by the. patient and other for hospital related. number of hospital beds and situated in the. characteristics.. non-West regions.. The regression results of using hospital. These results are consistent with much of. admissions for CMC as the dependent variable. the research on the determinants of access to. were strikingly similar to those using hospital. care [13-15]. For example, the finding that. admissions for ACSC as the dependent. blacks were more likely to have preventable. variable.. regression. hospitalization than whites is consistent with. coefficients indicate that most individual and. the finding that after adjustment for age and. hospital characteristics were significantly. health status, blacks had significantly fewer. associated with the dependent variables in the. ambulatory visits than their white counter-. same direction. The only exception for the. parts [16]. The linkage between access to care. patient variable was insurance status.. and hospital admissions for CMC indicates that. Workmen compensation was significant in the. the study of variations in hospitalization rates. ACSC model but not in the CMC model. The. for CMC can be used as an alternative measure. only exception for the hospital-related variable. of health care access. Our results confirmed. was ownership. Government ownership was. that certain demographic characteristics and. significant in the ACSC model but not in the. socioeconomic disadvantage are significant. CMC model.. barriers to the receipt of appropriate health. The. standardized. services [17]. Factors significantly associated with higher hospitalization rates for CMC can DISCUSSION. be used to develop a profile of individuals. Based on 1994 National Hospital Discharge. experiencing access barriers to primary and. Data, the results of this study showed that. ambulatory care. Policies and programs can. 11.7% of hospitalized adult patients younger. then be developed that aim at improving. than 65 years were diagnosed with at least. access to care for these individuals.. one of the five CMC studied (5.3% based on. The results of analyses based on hospital. principal diagnosis and 6.4% based on. admissions for CMC were consistent with. secondary diagnosis). Individual and hospital. analyses based on hospital admissions for. characteristics significantly associated with. ACSC as identified by Billing and associates. hospital admissions for CMC (objective 1). (second objective)[4]. By and large, the same. included age, gender, race, marital status. sets of individual and hospital characteristics. (individual predisposing factors), principal and. were associated with preventable hospi-. secondary sources of payment (individual. talizations whether measured by hospital. enabling factors), length of stay (individual. admissions for CMC or ACSC. The results are. need factor), number of hospital beds, and. also consistent with previous research on. geographic region (system factors). Hospital. hospitalization for ACSC. For example, the. ownership was not a significant predictor of. finding that the uninsured and Medicaid. preventable hospitalization. Specifically,. patients were more likely to be admitted for. individuals likely to be admitted with CMC. CMC. than. privately. insured. patients.

(11) Cheng-Chieh Lin, et al.. 19. corroborates the study by Weissman et al.. The cross-sectional nature of the data did not. with data from Massachusetts and Maryland. provide definitive conclusions about the. that examined the predictors of ACSC. specific causes associated with preventable. hospitalization [3]. To the extent that insurance. hospitalizations. A longitudinal or case-control. status serves as a proxy for income, the results. design would provide more valid conclusions.. of our analyses are also consistent with those. Furthermore, hospital admission rates for CMC. of Billings and his associates [1,4] based on. alone are not sufficient proof that the. New York data, and Bindman and his. provision of ambulatory care is inadequate. colleagues [6] using California data.. since other factors might also affect pre-. The consistent findings based on. ventable hospitalization including variations in. admissions for CMC and ACSC suggest that. disease prevalence, health care seeking. using hospital admissions for CMC can serve. behaviors, physician practice styles, and. as an efficient way of identifying subpopu-. system characteristics [6,14]. However, for areas. lations facing access barriers. The wide. with consistently high preventable hospi-. availability of hospital discharge data makes it. talization rates, we can be confident that. easy and convenient to calculate preventable. problems exist with the provision of. hospitalization rates. The analyses can be used. ambulatory care. Finally, the NHDS did not. to monitor and assess the effectiveness of. contain individual identifiers that could be. programs. at. used to locate repeated hospitalizations for the. improving access to care. Trend analyses can. and. interventions. aimed. same patients. To the extent there were. be conducted to measure progress over time.. systematic differences in readmission rates. Analyses can also be conducted for com-. across population groups especially between. parisons across communities or health plans.. the chronic medical users and others,. Preventable hospitalization rates can be. estimation biases were likely to occur.. conveniently incorporated in community health needs assessment and health plan quality report cards. Caution, however, must be exercised in conducting the analyses. Our. REFERENCES 1.. socioeconomic status on hospital use in New York. study indicates that individual sociodemographic characteristics significantly affect the preventable hospitalization rates and must be. City. Health Aff (Millwood) 1993;12:162-73. 2.. economic status and use of hospital resources.. One may argue that the measure of CMC. N Engl J Med 1990;322:1122-8. 3.. more specifically examine the independent. Weissman JS, Gatsonis C, Epstein AM. Rates of avoidable hospitalization by insurance status in. gross. More specific analyses should be related to particular diagnoses. Such analyses could. Epstein AM, Stern RS, Weissman JS. Do the poor cost more? A multihospital study of patients’ socio-. included in the analyses. including diverse diagnoses might be too. Billings JB, Zeitel L, Lukomnik J, et al. Impact of. Massachusetts and Maryland. JAMA 1992;268:2388-94. 4.. Billings JB, Zeitel L, Lukomnik J, et al. Analysis of variation in hospital admission rates associated with. effects of a particular diagnosis, and might. area income in New York City. NY, NY: United. provide more complete and understandable. Hospital Fund of New York, 1994.. results. While such explicit analyses are, in many ways, likely to be more informative, the more global analyses still have a role to play.. 5.. Caper P. The microanatomy of health care. Health. 6.. Bindman AB, Grumbach K, Osmond D, et al.. Aff (Millwood) 1993;12:174-7. Preventable hospitalizations and access to health. For example, global measures provide needed comprehensive indicators of the overall effects. care. JAMA 1995;274:305-11. 7.. Revision, Clinical Modification. CPHA Edwards Bro,. of CMC to inform national health policy makers. This study had a number of limitations.. International Classification of Disease, Ninth Ann Arbor, Mich, 1986.. 8.. National Center for Health Statistics. Development of the Design of the NCHS Hospital Discharge.

(12) Individual and Hospital Factors for Chronic Medical Conditions. 20. Survey. Vital and Health Statistics. PHS Pub. No. 1000, Series 2-No. 39. Public Health Service. Washington DC: US Government Printing Office, September 9.. PR, eds. Introduction to Health Services. Albany, NY: Delmar Publishers, 1993:46-70. 14. Long MJ. The Medical Care System: a Conceptual. 1970.. Model. Ann Arbor, MI:AUPHA Press/Health. SMG Hospital Marketing Group, Inc. Hospital. Administration Press, 1994.. Market Database. Healthcare Information Specialists, 1342 North LaSalle Drive, Chicago, Illinois, 1989. 10. Aday LA, Andersen RM. Equity of access to medical. 15. Ricketts TC. Barriers to Access to Services Provided by Physicians in General/Family Practice, General Internal Medicine, General Pediatrics, General. care: a conceptual and empirical overview. Med. Surgery, Obstetrics/Gynecology, and General and. Care 1981;19(Suppl): 4-27.. Child Psychiatry. Paper prepared for the Council on. 11. Mechanic D. Correlates of physician utilization: why. Graduate Medical Education. Cecil G. Sheps Center. do major multivariate studies of physician uti-. for Health Services Research, University of North. lization find trivial psychosocial and organizational effects? J Health Soc Behav 1979;20:387-96. 12. Markides KS, Levin JS, Ray LA. Determinants of physician utilization among Mexican-Americans: a three-generations study. Med Care 1985;23:236-46. 13. Aday LA. Indicators and predictors of health services utilization. In: Williams ST and Torrens. Carolina at Chapel Hill, August, 1991. 16. Kleinman JC, Gold M, Makuc D. Use of ambulatory medical care by the poor: another look at equity. Med Care 1981;19:1011-29. 17. Casanova C, Starfield B. Hospitalizations of children and access to primary care: a cross-national comparison. Int J Health Serv 1995;25:283-94..

(13) 21. 1. 2. 1,3 1. 3 2. ( ) (ambulatory care sensitivity conditions). 1994. 1999 21. 2 8/13/1998 11/2/1998. 10/1/1998. 4. 9-.

(14)

數據

Table 2 provides the operational definitions of the measures used in the analysis.
Table 2 provides the definitions and descriptive statistics of the variables used in our study

參考文獻

相關文件

because Nakagami parameter has physical meanings associated with envelope distribution and scatterer properties.  We used blue for pre-Rayleigh, white for Rayleigh, and red

Center for Faculty Development, Department of medical education, Taichung Veterans General hospital 1 School of Medicine, National Yang-Ming University, Taipei, Taiwan 2.

67 Número de médicos dos cuidados de saúde diferenciados, segundo o sexo, por especialidade Number of doctors in hospital care by gender and

67 Número de médicos dos cuidados de saúde diferenciados, segundo o sexo, por especialidade Number of doctors in hospital care by gender and

We have made a survey for the properties of SOC complementarity functions and theoretical results of related solution methods, including the merit function methods, the

We have made a survey for the properties of SOC complementarity functions and the- oretical results of related solution methods, including the merit function methods, the

[1] F. Goldfarb, Second-order cone programming, Math. Alzalg, M.Pirhaji, Elliptic cone optimization and prime-dual path-following algorithms, Optimization. Terlaky, Notes on duality

(i) spectral factorization associated with circular cones; (ii) smooth and nonsmooth analysis for f L θ given as in (7); (iii) the so-called L θ -convexity; and (iv) L θ