FACTORS ASSOCIATED WITH FREQUENT USE OF EMERGENCY SERVICES IN A MEDICAL CENTER

Transcription

1 FACTORS ASSOCIATED WITH FREQUENT USE OF EMERGENCY SERVICES IN A MEDICAL CENTER Jin-An Huang, 1 Wen-Chen Tsai, 2 Yin-Chieh Chen, 2 Wei-Hsiung Hu, 1 and Dar-Yu Yang 1 Background and Purpose: Overuse of emergency department (ED) services can result in overcrowding and a substantial increase of health care cost. The purpose of this study was to characterize frequent ED users and to identify the factors associated with frequent ED use in a hospital in Taiwan. Methods: This retrospective study used Andersen s Behavioral Model of Health Service Use as a framework. Frequent ED users ( 4 visits/year) and infrequent ED users (< 4 visits/ year) were selected randomly from patients visiting the adult ED of a medical center in central Taiwan from October 1, 2000 to September 30, Telephone interviews were completed for 200 frequent users and 600 infrequent users. Logistic regression analysis was performed to identify factors associated with frequent ED use. Results: Frequent ED users comprised 3.5% of total ED patients, accounting for 14.3% of the ED visits. Significant factors associated with frequent ED use were: a regular source of care [odds ratio (OR), 2.79; 95% confidence interval (CI), 1.63 to 4.79], alcoholism (OR, 19.4; 95% CI, 3.84 to 98.0), high outpatient clinic use (OR, 2.66; 95% CI, 1.72 to 4.11), previous hospitalization (OR, 3.06; 95% CI, 1.94 to 4.82), chronic disease (OR, 3.07; 95% CI, 1.78 to 5.29), cancer (OR, 4.16; 95% CI, 1.29 to 13.4), gastrointestinal disease (OR, 6.28; 95% CI, 1.95 to 20.2), cardiovascular disease (OR, 8.41; 95% CI, 2.51 to 28.1), pulmonary disease (OR, 4.21; 95% CI, 1.04 to 17.1), low level of emergency (OR, 5.43; 95% CI, 3.40 to 8.68), and dissatisfaction with treatment outcome (OR, 2.62; 95% CI, 1.32 to 5.20). Conclusions: Frequent ED users were responsible for a disproportionate number of the total ED visits. Andersen s need factors were strongly associated with frequent ED use, while the investigated predisposing factors were not significant. Most patients who visited the ED had a low level of emergency. These findings suggest that implementation of an integrated delivery system may decrease ED use. Key words: Emergency services, hospital; Health services, misuse; Health services, research; Risk factors J Formos Med Assoc 2003;102:222-8 Emergency departments (EDs) are being used with increasing frequency, 1 4 and a large portion of this increase is attributed to non-urgent or inappropriate visits. 5 8 Overuse of EDs is considered to result in overcrowding 5,9,10 and to contribute substantially to increased health care costs. 9,10 Overcrowding of EDs has become a significant problem in developed countries 11,12 and is associated with decreased quality of care, including poor outcome, prolonged suffering, long waits, dissatisfaction, staff frustration, and violence. 13,14 The rising cost of emergency medical service is becoming a major concern in Taiwan. In an increasingly cost-conscious health delivery system, inappropriate or non-urgent visits to the ED can be handled more efficiently and at a reduced cost by primary care. 15 Several western studies have demonstrated that a segment of the population uses the ED frequently and constitutes a considerable proportion of the total visits Frequent ED use has been associated with a lack of primary care and complex medical and social problems, 19 such as homelessness and alcoholrelated problems. Three distinctive patient populations were identified as potentially overusing EDs: public aid/uninsured patients, elderly patients, and pediatric patients. 20 No such study of the characteristics of individuals overusing EDs has been conducted in Taiwan. The purpose of this study was to determine the characteristics of frequent ED users, to compare frequent ED users with infrequent ED users, and to determine the factors associated with frequent ED use. 1 Department of Emergency Medicine, Taichung Veterans General Hospital, Taichung; 2 Institute of Health Services Management, China Medical College, Taichung, Taiwan. Received: 15 November 2002 Revised: 19 December 2002 Accepted: 11 February 2003 Reprint requests and correspondence to: Dr. Wen-Chen Tsai, Institute of Health Service Management, China Medical College, 91 Hsueh-Shih Road, Taichung, Taiwan. 222 J Formos Med Assoc 2003 Vol 102 No 4

2 Factors Associated with Frequent ED Use Methods Study design This study was based on Andersen s Behavioral Model of Health Services Use. 21 This model assumes that people s use of health services is a function of their predisposition to use services, factors for enabling use, and their need for use. Among the predisposing characteristics in Andersen s model, demographic factors (gender, age, and marital status) and social structure (level of education and employment status) were included in this study. Enabling resources were obtained for average monthly household income, financial barriers, the distance from hospital, and a regular source of care. Need factors included in this study comprised the following 9 components: subjective health status, alcoholism, high outpatient clinic use (> 24 visits/year, including physician visits in primary health care and hospital outpatient visits), hospital admission (in-hospital care), chronic disease (as designated by the Bureau of National Health Insurance in Taiwan), the most frequent diagnostic group during the successive ED visits, average triage status (high level of emergency as average triage level 1 or 2/low level of emergency as average triage level 3 or 4), average length of ED stay, and reason for ED use. Finally, we added patient satisfaction with the outcome of ED treatment to the list of included need factors. Subjects We conducted the study at Taichung Veterans General Hospital (TCVGH), a public teaching tertiary hospital in central Taiwan, located in a suburban area of Taichung. This hospital has approximately 55,000 visits per year. Any patient who visited the adult (15 years and older) ED from October 1, 2000 to September 30, 2001, was eligible for the study. A frequent ED user was defined as any patient who used the TCVGH ED 4 times or more in the 1-year study period. There were 1096 frequent users, and 29,890 infrequent users. Source of data Information concerning gender, age, chronic disease, diagnostic group, triage status, and length of ED stay was compiled retrospectively from the hospital information system database. One senior emergency physician reviewed the computer-based medical records for diagnostic group and alcoholism. Using a structured questionnaire, a telephone interview was given by using a computer-assisted telephone interview (CATI) system. With the aid of the CATI system, patients were randomly selected and interviewed by specialized investigators. The frequent ED users were interviewed until 200 questionnaires were completed. A total of 701 patients were called. The response rate was 28.5%. Twenty one percent of the patients called had died and were excluded from the analysis. Among the infrequent users, 600 questionnaires were completed for comparison. A total of 2631 patients were called. The response rate was 22.8%; 6.2% of the patients called had died and were excluded from the analysis. The survey was conducted between 9 am and 8 pm, from January 18 to February 5, Statistical analysis All data were managed and analyzed using Statistical Package for the Social Sciences (SPSS) software for Windows (version 10.0, SSPS Inc., Chicago, IL, USA). A frequency distribution was used to describe the demographic characteristics and the distribution of each variable. Chi-squared test was used to determine the differences between frequent ED users and infrequent ED users with regard to each variable. To test the likelihood of frequent ED use, a multiple logistic regression analysis was used with the forward stepwise method, including the likelihood-ratio test for the overall model, Hosmer-Lemeshow model goodness-of-fit analysis, overall percentage correctly predicted, and odds ratios (ORs) with 95% confidence intervals (CIs). A p value of < 0.05 was considered statistically significant. Results Frequent ED users comprised only 3.5% of total patients, but accounted for 14.3% of all visits to the ED. The 10 patients with the greatest number of visits to the ED visited 12 to 23 times each, averaging 16.2 visits. Tables 1 and 2 show the patient characteristics and differences between the frequent and infrequent ED users. Within the frequent ED user group, the proportion of men was insignificantly higher than that of women (χ 2 = 3.155, p = 0.076). There was also no significant difference in distance from the hospital between the frequent and infrequent ED users (χ 2 = 2.424, p = 0.119). All other 17 variables were significantly different between the frequent users and the infrequent users (chi-squared test, p < 0.05), and were entered into a multiple logistic regression analysis that used forward selection with a likelihood ratio test for backward elimination. The logistic regression model for frequent ED use is shown in Table 3. The correlation matrix indicated that there were no high correlations between the explanatory variables in the model. In addition, no significant collinearity between the explanatory J Formos Med Assoc 2003 Vol 102 No 4 223

3 Table 1. Differences in patient characteristics between infrequent and frequent emergency department users. Variable Infrequent users Frequent users p Value* (n = 600) [n (%)] (n = 200) [n (%)] Predisposing characteristics Gender Female 265 (44.2) 74 (37.0) Male 335 (55.8) 126 (63.0) Age (years) < (34.5) 28 (14.0) (35.0) 79 (39.5) > (30.5) 93 (46.5) Marital status Married 118 (19.7) 21 (10.5) Single 428 (71.3) 156 (78.0) Others 54 (9.0) 23 (11.5) Level of education Less than high school 232 (38.7) 103 (51.5) High school completed 197 (32.8) 63 (31.5) College or higher 171 (28.5) 34 (17.0) Employment status Working 283 (47.2) 72 (36.0) Not working 317 (52.8) 128 (64.0) Enabling resources Monthly household income (NT$) < 30, (48.5) 128 (64.0) 30,000 59, (30.0) 43 (21.5) 60,000 89, (10.7) 12 (6.0) 90, (10.8) 17 (8.5) Financial barrier No 493 (82.2) 144 (72.0) Yes 107 (17.8) 56 (28.0) Distance from hospital < 30 min 265 (44.2) 101 (50.5) 30 min 335 (55.8) 99 (49.5) Regular source of care < No 216 (36.0) 28 (14.0) Yes 384 (64.0) 172 (86.0) * Pearson s chi-squared test. variables was revealed by the tolerance statistics and the variance inflation factor (VIF). No predisposing factor (age, marital status, education, and employment status) was significantly related to frequent ED use. Of the enabling factors, patients with a regular source of care were more likely to be frequent ED users (OR, 2.79; 95% CI, 1.63 to 4.79). Neither the average monthly household income nor financial barriers had any significant effect on frequent ED use. Predictably, the most important factors related to frequent ED use appeared to be related to the need factors. Patients with potential for frequent ED use were those with alcohol dependence (OR, 19.4; 95% CI, 3.84 to 98.0); high outpatient clinic users (OR, 2.66; 95% CI, 1.72 to 4.11); those who had previously used in-hospital care (OR, 3.06; 95% CI, 1.94 to 4.82); those who had a chronic disease (OR, 3.07; 95% CI, 1.78 to 5.29); those who had cancer (OR, 4.16; 95% CI 1.29 to 13.4), gastrointestinal disease (OR, 6.28; 95% CI, 1.95 to 20.2), cardiovascular disease (OR, 8.41; 95% CI, 2.51 to 28.1) or pulmonary disease (OR, 4.21; 95% CI, 1.04 to 17.1); and those who had a low level of emergency (OR, 5.43; 95% CI, 3.40 to 8.68). Need factors not significantly related to frequent ED use were poor subjective health status, a longer stay in the ED, and reasons for ED use. Patients who were dissatisfied with care on discharge from the ED were also more likely to use the ED frequently (OR, 2.62; 95% CI, 1.32 to 5.20). Discussion Survey data in this study showed that frequent ED users comprised a relatively small number of patients who were responsible for a disproportionate number of total ED visits. The finding that 3.5% of the patients at a hospital ED accounted for 14.3% of the visits in our study is similar to the finding of studies from western countries that defined frequent ED use as 4 or more visits per patient per year. 224 J Formos Med Assoc 2003 Vol 102 No 4

4 Table 2. Differences in needs factors between infrequent and frequent emergency department (ED) users. Factors Associated with Frequent ED Use Variable Infrequent users Frequent users p Value* (n = 600) [n (%)] (n = 200) [n (%)] Subjective health status < Poor 168 (28.0) 96 (48.0) Average 259 (43.2) 73 (36.5) Good 173 (28.8) 31 (15.5) Alcoholism < No 597 (99.5) 189 (94.5) Yes 3 (0.5) 11 (5.5) High outpatient clinic use < No 469 (85.6) 79 (14.4) Yes 131 (52.0) 121 (48.0) In-hospital care < None 348 (58.0) 44 (22.0) Yes 252 (42.0) 156 (78.0) Chronic disease < None 282 (47.0) 31 (15.5) Yes 318 (53.0) 169 (84.5) Diagnostic groups < Trauma 94 (15.7) 5 (2.5) Cancer 33 (5.5) 39 (19.5) Gastrointestinal disease 49 (8.2) 29 (14.5) Cardiovascular disease 32 (5.3) 22 (11.0) Genitourinary disease 35 (5.8) 17 (8.5) Psychiatric disease 25 (4.2) 14 (7.0) Pulmonary disease 20 (3.3) 10 (5.0) Obstetrics and gynecology 14 (2.3) 4 (2.0) Acute upper respiratory tract infection 34 (5.7) 3 (1.5) Non-specific abdominal pain 28 (4.7) 3 (1.5) Others 236 (39.3) 54 (27.0) Average triage status < High level of emergency 369 (61.5) 58 (29.0) Low level of emergency 231 (38.5) 142 (71.0) Average length of ED stay < 48 hours 543 (90.5) 171 (85.5) 48 hours 57 (9.5) 29 (14.5) Reason for ED Doctor s suggestion 47 (7.8) 11 (5.5) ED is faster 25 (4.2) 8 (4.0) Problem was serious 464 (77.3) 174 (87.0) Others 64 (10.7) 7 (3.5) Satisfaction with treatment outcome < Poor 45 (7.5) 32 (16.0) Fair 205 (34.2) 114 (57.0) Good 350 (58.3) 54 (27.0) * Pearson s chi-squared test. As expected, the factors most strongly related to frequent ED use in our model were the indicators of health needs. Review of clinical medical records revealed that frequent ED users showed high rates of multiple chronic medical conditions, including alcohol abuse and psychiatric illness, which is in agreement with previous studies. 25,26 Because patients may have ill-defined conditions or have different reasons for each visit, registering a single diagnosis is difficult. However, we categorized patients by the most frequent diagnoses for patients with successive visits during the study year. Cancers were the single most frequent definitive diagnosis related to frequent ED use, followed by gastrointestinal diseases and cardiovascular diseases. Cancer has been the leading cause of death in Taiwan since 1982, accounting for nearly 1 out of 4 deaths. It is quite obvious that acute care may not be appropriate for terminal cancer patients. Rather, by taking all needs into consideration, hospice care provides a better quality of life for these patients. 27 In contrast to the rapid growth of the hospice movement in the developed countries such as the USA, the UK, and Japan, hospice service has had a low growth rate in Taiwan. 28 Owing to the lack of availability of both hospital-based and home-based hospice care, cancer J Formos Med Assoc 2003 Vol 102 No 4 225

5 Table 3. Logistic regression model for frequent emergency department use (n = 800).*,,, Variable OR 95% CI Regular source of care No Yes Alcoholism No Yes High outpatient clinic use No Yes In-hospital care None Yes Chronic disease None Yes Diagnostic group Cancer Gastrointestinal disease Cardiovascular disease Pulmonary disease Average triage status High level of emergency Low level of emergency Satisfaction with treatment outcome Poor Fair Good - - * Model chi-squared test: χ 2 = , p < Goodness-of-fit test: χ 2 = 3.946, p = Cox and Snell R 2 = 0.334; Nagelkerke R 2 = Overall percentage correctly predicted: 83.6%. Reference group: trauma. OR = odds ratio; 95% CI = 95% confidence interval. patients might be forced to rely on emergency service care and in-hospital care instead. Emergency service care covers a higher proportion of total ambulatory care in some geographic areas, even in a health care system with an adequate supply of primary care physicians and universal insurance. 6 In addition, a subgroup of the population uses the hospital ED as a regular source of care. 29 These factors may explain why the increasing frequency of ED visits is not only associated with increasing percentages of patients who use other health care services besides the ED, but is also associated with increasing frequency of care at other sites. 22 Similarly, as was found in this study, patients who were heavy users of ambulatory care other than EDs were approximately 2.5 times as likely to be frequent ED users. Patients who had any previous use of in-hospital care were also about 3 times as likely to be frequent ED users. That is, high use of health care services other than the ED is strongly associated with frequent use of hospital EDs. Most ED patients, whether frequent ED users or not, thought that their problems were serious and needed immediate medical attention. However, it has been shown that a large portion of ED visits can be categorized as inappropriate or non-urgent use. 5,7 9 The increasing use of the ED is not associated with a change in the proportion of severe cases being seen. 3 Our study demonstrated that patients who visited ED with low level of emergency were more likely be frequent ED users. Patients who frequently seek medical attention at busy or overcrowded EDs for complaints judged by the triage staff to be not emergent, are inevitably given low priority. Olsson and Hansagi have shown that satisfaction with emergency service care is adversely affected by the patient perception that the ED staff have classified their use of the ED as inappropriate or belittled their symptoms. 18 Apart from the risk of overlooking true health emergencies, regarding the conditions of frequent ED users as non-urgent may result in long waiting times and dissatisfied patients. 22 To make matters worse, dissatisfaction is significantly related to increased utilization of emergency services care. In Taiwan, the National Health Insurance is a monopoly and its coverage has reached more than 96% of the population since its implementation in Because the fixed fee for ED care is just US$12 and most patients think their condition is serious, financial barriers are unlikely to underlie the frequent use of the ED. In addition, hospital-based care and primary care have been largely separated from each other in Taiwan, even though a global budget payment system was implemented in Most hospitals have a large outpatient department (OPD). Many patients may visit the ED or OPD in a teaching tertiary hospital merely for a common cold. Andersen showed that patients who had a regular source of care might use health care services more frequently. 21 Thus, patients who had a regular source of care were more likely to be frequent ED users. In this study, age, marital status, level of occupation, and employment status were significantly different between the frequent ED users and infrequent ED users (χ 2 test, p < 0.05). However, no predisposing factor included in the logistic model predicted frequent use of the ED. In Taiwan, older patients tend to have a lower level of education and are likely not to be working. Our analysis showed a high correlation of age and chronic disease. Consequently, frequent ED use in the elderly was significantly correlated with the presence of a chronic disease. Hospitals are now providing post-hospital disease management programs for selected chronic conditions that account for a high volume of repeat ED visits, such as cancer and chronic heart failure. 30 Disease management is a target in cost containment 226 J Formos Med Assoc 2003 Vol 102 No 4

6 Factors Associated with Frequent ED Use efforts. Case management has been shown to be a promising and cost-effective means of decreasing visits among frequent ED users. 19,31 This study had several limitations. The sampled hospital is a suburban, tertiary public teaching medical center. The proportion of patients with emergent triage status in the study population may thus have been higher than for other EDs. In addition, the demographic characteristics of the patients may not be representative of patients in other parts of the country. Similarly, infrequent ED users in the sampled hospital may have been frequent ED users in other hospitals. Although the overall response rate for the survey was more than 20%, interviews could not be held for patients who had died or who could not be contacted by telephone. Moreover, 21% of the subjects in the frequent ED group had died, a much higher percentage than that of the infrequent ED group (6.2%). Such differences may have biased the findings. A recall bias may be assumed because of the long lag between ED use and the telephone interview. Furthermore, conducting telephone interviews from 9 am to 8 pm may also have created a selection bias. In summary, frequent ED users accounted for a disproportionate number of the total ED visits. The need factors were strongly associated with frequent use of the ED, while the predisposing factors were not significant in this regard. However, most patients visited the ED with a low level of emergency. Cancer was the leading diagnosis of the frequent ED users. Development of an integrated cancer management program in Taiwan, including both hospital-based and home-based hospice care, may reduce unnecessary use of the ED in these patients. Efforts to facilitate closer contact between general practitioners, the ED, and other hospital departments may also improve quality of care and decrease the number of inappropriate ED visits. ACKNOWLEDGMENT: This study was supported by a grant (TCVGH A) from Taichung Veterans General Hospital, Taichung, Taiwan. References 1. Burt CW, McCaig LF: Trends in hospital emergency department utilization: United States, Vital Health Stat : Capewell S: The continuing rise in emergency admissions. BMJ 1996;312: Santos Eggimann B: Increasing use of the emergency department in a Swiss hospital: observational study based on measures of the severity of cases. BMJ 2002;324: Department of Health: Health and National Health Insurance Annual Statistics. Taipei: DOH, R.O.C; Hu SC: Clinical and demographic characteristics of 13,911 medical emergency patients. J Formos Med Assoc 1991;90: Mustard CA, Kozyrskyj AL, Barer ML, et al: Emergency department use as a component of total ambulatory care: a population perspective. CMAJ 1998;158: Pereira S, Oliveira e Silva A, Quintas M, et al: Appropriateness of emergency department visits in a Portuguese university hospital. Ann Emerg Med 2001;37: Sempere Selva T, Peiro S, Sendra Pina P, et al: Inappropriate use of an accident and emergency department: magnitude, associated factors, and reasons an approach with explicit criteria. Ann Emerg Med 2001;37: Kellermann AL: Nonurgent emergency department visits. Meeting an unmet need. JAMA 1994;271: Baker DW, Stevens CD, Brook RH: Patients who leave a public hospital emergency department without being seen by a physician. Causes and consequences. JAMA 1991;266: Shih FY, Ma MH, Chen SC, et al: ED overcrowding in Taiwan: facts and strategies. Am J Emerg Med 1999;17: Derlet R, Richards J, Kravitz R: Frequent overcrowding in U.S. emergency departments. Acad Emerg Med 2001;8: Miro O, Antonio MT, Jimenez S, et al: Decreased health care quality associated with emergency department overcrowding. Eur J Emerg Med 1999;6: Derlet RW, Richards JR: Overcrowding in the nation s emergency departments: complex causes and disturbing effects. Ann Emerg Med 2000;35: Glick DF, Thompson KM: Analysis of emergency room use for primary care needs. Nurs Econ 1997;15: Lucas RH, Sanford SM: An analysis of frequent users of emergency care at an urban university hospital. Ann Emerg Med 1998;32: Okuyemi KS, Frey B: Describing and predicting frequent users of an emergency department. J Assoc Acad Minor Phys 2001; 12: Olsson M, Hansagi H: Repeated use of the emergency department: qualitative study of the patient s perspective. Emerg Med J 2001;18: Pope D, Fernandes CM, Bouthillette F, et al: Frequent users of the emergency department: a program to improve care and reduce visits. CMAJ 2000;162: Bond TK, Stearns S, Peters M: Analysis of chronic emergency department use. Nurs Econ 1999;17:207-13, Andersen RM: Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav 1995;36: Hansagi H, Olsson M, Sjoberg S, et al: Frequent use of the hospital emergency department is indicative of high use of other health care services. Ann Emerg Med 2001;37: Murphy AW, Leonard C, Plunkett PK, et al: Characteristics of attenders and their attendances at an urban accident and emergency department over a one year period. J Accid Emerg Med 1999;16: Malone RE: Heavy users of emergency services: social construction of a policy problem. Soc Sci Med 1995;40: J Formos Med Assoc 2003 Vol 102 No 4 227

7 25. Hansagi H, Allebeck P, Edhag O, et al: Frequency of emergency department attendances as a predictor of mortality: nine-year follow-up of a population-based cohort. J Public Health Med 1990;12: Genell AK, Rosenqvist U: Heavy users of an emergency department a two year follow-up study. Soc Sci Med 1987; 25: Greer DS, Mor V, Morris JN, et al: An alternative in terminal care: results of the National Hospice Study. J Chronic Dis 1986; 39: Lo JC: The impact of hospices on health care expenditures the case of Taiwan. Soc Sci Med 2002;54: Lang T, Davido A, Diakite B, et al: Using the hospital emergency department as a regular source of care. Eur J Epidemiol 1997; 13: Goldstein R: The disease management approach to cost containment. Nurs Case Manag 1998;3: Okin RL, Boccellari A, Azocar F, et al: The effects of clinical case management on hospital service use among ED frequent users. Am J Emerg Med 2000;18: J Formos Med Assoc 2003 Vol 102 No 4