CHRONIC DISEASE
Chronic obstructive pulmonary disease hospitalisations and mortality in Victoria: analysis of variations by socioeconomic status Timothy Ore,1 Paul Ireland1
S
ocioeconomic status (SES) is a key determinant of health outcomes, particularly for chronic diseases.1-8 In a systematic review of chronic obstructive pulmonary disease (COPD) and SES, Gershon et al.1 found that individuals at the lowest socioeconomic strata were at least twice as likely to have poor outcomes as those of the highest, regardless of gender, age and population. The differences are associated with multiple factors, including variations in access to health care,6,9 diagnostic practices,6 cigarette smoking,10-11 occupational and environmental exposures,2,12-18 childhood respiratory infections17,19-20 and pre-term births.21
Abstract Objective: This paper analysed chronic obstructive pulmonary disease (COPD) hospitalisations, unplanned readmissions and deaths in Victoria to identify associations with socioeconomic status (SES). Methods: The data was taken from the Victorian Admitted Episodes Dataset, the Victorian Health Information Surveillance System, the Victorian Burden of Disease Study and the Australian Bureau of Statistics’ Index of Relative Socioeconomic Disadvantage. Results: COPD separations have a greater variation by SES than all separations. The average age-standardised separation rate (10.43) for the top percentile Local Government Areas (LGA) was 5.8 times that of the bottom percentile LGAs (1.80). The top percentile group was the lowest SES group (effect size = 0.93). There were significant negative correlations between the age-standardised COPD separation rates and SES across LGAs (r = -0.60) and Regions (r = -0.89). Analysis of readmissions (r = -0.49), mortality data (r = -0.51) and the burden of disease data (r = -0.39) also showed significant inverse associations between COPD and SES.
In an analysis of data from the 2011 Behavioural Risk Factor Surveillance System in the US,5 respondents who did not have a high school diploma reported a higher prevalence of COPD (9.5%) than those with a high school diploma (6.8%) or some college (4.6%). Reported COPD prevalence decreased with increasing household income, more current smokers reported having COPD (13%) than former smokers (6.8%) or never smokers (2.8%), and respondents with a history of asthma were significantly more likely to have been diagnosed with COPD (20.3%) than those without asthma (3.8%). In the US, occupations that had the highest risk for COPD (smokers and never smokers combined) were freight, stock and material handlers, the armed forces, vehicle mechanics and records processing and distribution clerks.22 Among never smokers, the risk was also high for machine operators and
Conclusions and implications: Victorians living in the most disadvantaged areas have a greater burden from COPD, highlighting a need to prioritise public health services interventions to improve outcomes. Key words: coefficient of variation, effect size, socioeconomic gradient, hospital readmissions, burden of chronic disease construction trades and labourers. In New Zealand, Fiswick et al.23 found an association between airflow obstruction and working as cleaners, bakers, spray painters, laboratory technicians, and plastics and rubber workers and in construction and mining. Chronic bronchitis was reported more often in food processors, chemical processors and spray painters and in construction and mining. COPD often results in hospitalisation, with systemic manifestations and physical comorbidities including depression,24 chest infections, coronary artery disease and right heart failure.25 Mental and physical health co-morbidities in COPD contribute to adverse
health outcomes, such as reduced adherence to management plans, slower recovery from exacerbations, increased health care utilisation and higher mortality rates.25 In 2010–11, Australia’s rate of COPD was 317 hospital separations per 100,000 population (aged 15 years and over), more than 56% higher than the Organisation for Economic Cooperation and Development (OECD) average.26 During 2007–08 and 2011–12, the average COPD separation rates did not change substantially in Australia, but the OECD average fell from 223 in 2006 to 203 in 2011.9 In the US, COPD was responsible for 8 million physician office visits, 1.5 million
1. Department of Health, Commission for Hospital Improvement, Victoria Correspondence to: Dr Timothy Ore, Department of Health, Commission for Hospital Improvement, Level 20, 50 Lonsdale Street, Melbourne, Victoria 3000; e-mail: [email protected] Submitted: June 2014; Revision requested: August 2014; Accepted: August 2014 The authors have stated they have no conflict of interest. Aust NZ J Public Health. 2015; 39:243-9; doi: 10.1111/1753-6405.12305
2015 vol. 39 no. 3
Australian and New Zealand Journal of Public Health © 2014 Public Health Association of Australia
243
Ore and Ireland
Article
emergency department visits and 727,000 hospitalisations (13% of all hospitalisations) in 2000.27 COPD is second only to coronary heart disease as a reason for payment of Social Security disability benefits in the US.27
area is compared with other areas. A high IRSED indicates that an area has few families of low income and few people with little training and in unskilled occupations.
This paper analysed COPD hospitalisation, unplanned readmission and death data across all 79 Local Government Areas (LGAs) in Victoria in 2011–12 to identify the direction and magnitude of associations between SES and COPD. Prevalence data reported in the 2001 Victorian Burden of Disease survey was also examined with a similar objective. This study, the first to specifically examine SES and COPD hospitalisations and readmissions at a detailed level in Victoria, points to a need for further actions on prevention, early diagnosis and effective treatment strategies to manage the burden from COPD. An understanding of the relationship is important for designing clinical and policy strategies to improve the health of people with COPD.1
Statistical analysis
Methods Data The numerator data (separations and 30-day unplanned readmissions) was from the Victorian Admitted Episodes Dataset, public and private hospitals, for 2011–12, for patients with a COPD diagnosis. COPD is a generic term for emphysema and chronic bronchitis. Death data (avoidable mortality) for 2002–06 was taken from the Victorian Health Information Surveillance System, containing untimely and unnecessary deaths from diseases for which effective public health and medical intervention are available. The denominator data (persons aged 25 years and over) was from the Estimated Resident Population (ERP) by the Australian Bureau of Statistics (ABS). The ERP is the official estimate of the Australian population, based on usual place of residence. The data is provided at several levels, including at the LGA. Data was also taken from the Victorian Burden of Disease Study: Mortality and Morbidity in 2001.28 This provides a comprehensive assessment of the health status of Victorians, including chronic respiratory disease. For socioeconomic status, the ABS Index of Relative Socioeconomic Disadvantage (IRSED) was used. The index is derived from attributes including low income, low educational attainment, high unemployment and jobs in relatively unskilled occupations. The higher an area’s IRSED, the less disadvantaged that
244
Rates were calculated as the number of COPD hospital separations (for patients over 25 years of age) divided by the estimated resident population for the corresponding age group, multiplied by 1,000. Readmission rates were per 10,000 population and death rates were per 100,000 population. Patients younger than 25 years of age represented less than 1% of the separations. The data was analysed at the LGA level. The rates were agestandardised using the direct method, with Victoria’s 2012 population as the reference. Statistical analyses were conducted using SPSS, version 20. The coefficient of variation (CV) was calculated for COPD separation and mortality rates, and compared with the distribution of rates for all separations and deaths by LGA. CV is computed by dividing the standard deviation of the data set by its mean. The distribution of all separation and death rates by LGA is a useful data set for comparison to assess the relative degree of variability between conditions. As noted by Refshauge and Kalisch,6 when a condition is common it is more likely to have a wide distribution of rates. Therefore, when comparing the variation in rates for conditions that differ in prevalence, it is desirable to adjust for this effect.
Results Profile There were 15,263 separations (including 4,082 with catastrophic complications or comorbidities) from 1 July 2011 to 30 June 2012 in Victoria (Table 1). The average length of stay of was 5.98 days. Patients younger than 25 years accounted for 0.4% (n=62) of the separations. The rates increased significantly (p
Chronic obstructive pulmonary disease hospitalisations and mortality in Victoria: analysis of variations by socioeconomic status Timothy Ore,1 Paul Ireland1
S
ocioeconomic status (SES) is a key determinant of health outcomes, particularly for chronic diseases.1-8 In a systematic review of chronic obstructive pulmonary disease (COPD) and SES, Gershon et al.1 found that individuals at the lowest socioeconomic strata were at least twice as likely to have poor outcomes as those of the highest, regardless of gender, age and population. The differences are associated with multiple factors, including variations in access to health care,6,9 diagnostic practices,6 cigarette smoking,10-11 occupational and environmental exposures,2,12-18 childhood respiratory infections17,19-20 and pre-term births.21
Abstract Objective: This paper analysed chronic obstructive pulmonary disease (COPD) hospitalisations, unplanned readmissions and deaths in Victoria to identify associations with socioeconomic status (SES). Methods: The data was taken from the Victorian Admitted Episodes Dataset, the Victorian Health Information Surveillance System, the Victorian Burden of Disease Study and the Australian Bureau of Statistics’ Index of Relative Socioeconomic Disadvantage. Results: COPD separations have a greater variation by SES than all separations. The average age-standardised separation rate (10.43) for the top percentile Local Government Areas (LGA) was 5.8 times that of the bottom percentile LGAs (1.80). The top percentile group was the lowest SES group (effect size = 0.93). There were significant negative correlations between the age-standardised COPD separation rates and SES across LGAs (r = -0.60) and Regions (r = -0.89). Analysis of readmissions (r = -0.49), mortality data (r = -0.51) and the burden of disease data (r = -0.39) also showed significant inverse associations between COPD and SES.
In an analysis of data from the 2011 Behavioural Risk Factor Surveillance System in the US,5 respondents who did not have a high school diploma reported a higher prevalence of COPD (9.5%) than those with a high school diploma (6.8%) or some college (4.6%). Reported COPD prevalence decreased with increasing household income, more current smokers reported having COPD (13%) than former smokers (6.8%) or never smokers (2.8%), and respondents with a history of asthma were significantly more likely to have been diagnosed with COPD (20.3%) than those without asthma (3.8%). In the US, occupations that had the highest risk for COPD (smokers and never smokers combined) were freight, stock and material handlers, the armed forces, vehicle mechanics and records processing and distribution clerks.22 Among never smokers, the risk was also high for machine operators and
Conclusions and implications: Victorians living in the most disadvantaged areas have a greater burden from COPD, highlighting a need to prioritise public health services interventions to improve outcomes. Key words: coefficient of variation, effect size, socioeconomic gradient, hospital readmissions, burden of chronic disease construction trades and labourers. In New Zealand, Fiswick et al.23 found an association between airflow obstruction and working as cleaners, bakers, spray painters, laboratory technicians, and plastics and rubber workers and in construction and mining. Chronic bronchitis was reported more often in food processors, chemical processors and spray painters and in construction and mining. COPD often results in hospitalisation, with systemic manifestations and physical comorbidities including depression,24 chest infections, coronary artery disease and right heart failure.25 Mental and physical health co-morbidities in COPD contribute to adverse
health outcomes, such as reduced adherence to management plans, slower recovery from exacerbations, increased health care utilisation and higher mortality rates.25 In 2010–11, Australia’s rate of COPD was 317 hospital separations per 100,000 population (aged 15 years and over), more than 56% higher than the Organisation for Economic Cooperation and Development (OECD) average.26 During 2007–08 and 2011–12, the average COPD separation rates did not change substantially in Australia, but the OECD average fell from 223 in 2006 to 203 in 2011.9 In the US, COPD was responsible for 8 million physician office visits, 1.5 million
1. Department of Health, Commission for Hospital Improvement, Victoria Correspondence to: Dr Timothy Ore, Department of Health, Commission for Hospital Improvement, Level 20, 50 Lonsdale Street, Melbourne, Victoria 3000; e-mail: [email protected] Submitted: June 2014; Revision requested: August 2014; Accepted: August 2014 The authors have stated they have no conflict of interest. Aust NZ J Public Health. 2015; 39:243-9; doi: 10.1111/1753-6405.12305
2015 vol. 39 no. 3
Australian and New Zealand Journal of Public Health © 2014 Public Health Association of Australia
243
Ore and Ireland
Article
emergency department visits and 727,000 hospitalisations (13% of all hospitalisations) in 2000.27 COPD is second only to coronary heart disease as a reason for payment of Social Security disability benefits in the US.27
area is compared with other areas. A high IRSED indicates that an area has few families of low income and few people with little training and in unskilled occupations.
This paper analysed COPD hospitalisation, unplanned readmission and death data across all 79 Local Government Areas (LGAs) in Victoria in 2011–12 to identify the direction and magnitude of associations between SES and COPD. Prevalence data reported in the 2001 Victorian Burden of Disease survey was also examined with a similar objective. This study, the first to specifically examine SES and COPD hospitalisations and readmissions at a detailed level in Victoria, points to a need for further actions on prevention, early diagnosis and effective treatment strategies to manage the burden from COPD. An understanding of the relationship is important for designing clinical and policy strategies to improve the health of people with COPD.1
Statistical analysis
Methods Data The numerator data (separations and 30-day unplanned readmissions) was from the Victorian Admitted Episodes Dataset, public and private hospitals, for 2011–12, for patients with a COPD diagnosis. COPD is a generic term for emphysema and chronic bronchitis. Death data (avoidable mortality) for 2002–06 was taken from the Victorian Health Information Surveillance System, containing untimely and unnecessary deaths from diseases for which effective public health and medical intervention are available. The denominator data (persons aged 25 years and over) was from the Estimated Resident Population (ERP) by the Australian Bureau of Statistics (ABS). The ERP is the official estimate of the Australian population, based on usual place of residence. The data is provided at several levels, including at the LGA. Data was also taken from the Victorian Burden of Disease Study: Mortality and Morbidity in 2001.28 This provides a comprehensive assessment of the health status of Victorians, including chronic respiratory disease. For socioeconomic status, the ABS Index of Relative Socioeconomic Disadvantage (IRSED) was used. The index is derived from attributes including low income, low educational attainment, high unemployment and jobs in relatively unskilled occupations. The higher an area’s IRSED, the less disadvantaged that
244
Rates were calculated as the number of COPD hospital separations (for patients over 25 years of age) divided by the estimated resident population for the corresponding age group, multiplied by 1,000. Readmission rates were per 10,000 population and death rates were per 100,000 population. Patients younger than 25 years of age represented less than 1% of the separations. The data was analysed at the LGA level. The rates were agestandardised using the direct method, with Victoria’s 2012 population as the reference. Statistical analyses were conducted using SPSS, version 20. The coefficient of variation (CV) was calculated for COPD separation and mortality rates, and compared with the distribution of rates for all separations and deaths by LGA. CV is computed by dividing the standard deviation of the data set by its mean. The distribution of all separation and death rates by LGA is a useful data set for comparison to assess the relative degree of variability between conditions. As noted by Refshauge and Kalisch,6 when a condition is common it is more likely to have a wide distribution of rates. Therefore, when comparing the variation in rates for conditions that differ in prevalence, it is desirable to adjust for this effect.
Results Profile There were 15,263 separations (including 4,082 with catastrophic complications or comorbidities) from 1 July 2011 to 30 June 2012 in Victoria (Table 1). The average length of stay of was 5.98 days. Patients younger than 25 years accounted for 0.4% (n=62) of the separations. The rates increased significantly (p
