Research Meets Reality: Administrative Data to Guide Planning for Canadian Regional Health Authorities

The use of population-based information systems can help regional health authorities provide integrated healthcare. Administrative data from 1995 through 1997 for Manitoba's rural South Eastman Regional Health Authority and for the rest of Manitoba were used to aid South Eastman's strategic planning process. Three areas of concern were highlighted: the relatively high risk of poor health among residents of the RHA's Southern District, physician maldistribution and inconsistent patterns of service utilization. By improving population access to primary healthcare, regional imbalances are being addressed.     

Trends in socioeconomic health inequalities in Korea: use of mortality and morbidity measures

STUDY OBJECTIVE: To examine trends in educational mortality and morbidity inequalities in Korea. DESIGN: Census data (1990, 1995, 2000) and death certificate data (1990-91, 1995-96, 2000-01) were used for mortality. For morbidity, four waves (1989, 1992, 1995, and 1999) of Social Statistics Survey from Korea's National Statistical Office were used. Morbidity indicators were self rated health and self reported illness in the past two weeks. Trends were studied using indices for both the relative and absolute size of socioeconomic inequalities in health. SETTING: South Korea. Patients (or Participants): Representative annual samples of the adult population aged 30-59 in Korea. Main results: Based on trends in relative index of inequalities, the relative level of socioeconomic mortality inequality remained virtually unchanged in men and women in the past 10 years. Meanwhile, inequalities in self rated health have increased over time in both sexes. Most of the total increase in health inequalities happened between 1995 and 1999. Inequalities in self reported acute illness increased in the past 10 years. CONCLUSIONS: The rise in inequalities in morbidity requires increased social discourse and policy discussions about health inequalities in Korean society.     

Across time and space: variations in hospital use during Canadian health reform

OBJECTIVES: To investigate change in hospital utilization in a population and to discuss analytical strategies using large administrative databases, focusing on variations in rates of different types of hospital utilization by income quintile neighborhoods. DATA SOURCES: Hospital discharge abstracts from Manitoba Health, used to study the changes in utilization rates over eight fiscal years (1989-1996). STUDY DESIGN: We test the hypotheses that health reform has changed utilization rates, that utilization rates differ significantly across income quintiles (defined by the relative affluence of neighborhood of residence), and that these variations have been maintained over time. Our approach uses generalized estimating equations to produce robust and consistent results for studying rates of recurrent and nonrecurrent events longitudinally. DATA EXTRACTION METHODS: Rates of individuals hospitalized, hospital discharges, days of hospitalization, and hospitalization for different types of medical conditions and surgical procedures are generated for the period April 1, 1989 through March 31, 1997 for residents of Winnipeg, Manitoba. Data are grouped according to the individual's age, gender, and neighborhood of residence on April 1 of each of the eight fiscal years for the rate calculations. Neighborhood of residence and the 1991 Canadian Census public use database are used to assign individuals to income quintiles. PRINCIPAL FINDINGS: The substitution of outpatient surgery for inhospital surgery accounted for much of the change in hospital utilization over the 1989-1996 period. Health care reform did not have a significant effect on the utilization gradient already observed across socioeconomic groups. Health reform markedly accelerated declines in in-hospital utilization. CONCLUSIONS: Grouping the data with key characteristics intact facilitates the statistical analysis of utilization measures previously difficult to study. Such analyses of variations across time and space based on parametric models allows adjustment for continuous covariates and is more efficient than the traditional nonparametric approach using standardized rates.     

Geographic variation in primary care visits in Iowa.

OBJECTIVE. This study investigates the determinants of primary care office visit rates. DATA SOURCES. Blue Cross and Blue Shield of Iowa subscriber information was sorted by residence into geographic health service areas. Cost-sharing information was also obtained from Blue Cross. Physician supply data were obtained from The University of Iowa, Office of Community-Based Programs. Hospital data were reported by the Iowa Hospital Association. STUDY DESIGN. Cases were classified into ambulatory care groups (ACGs). Use rates were computed for each group in each service area. Ordinary least squares regression models were developed to model geographic variation in each ACG-specific primary care visit rate. PRINCIPAL FINDINGS. Regression models were not significant for five out of eleven ACGs studied. Out-of-pocket expense significantly affected utilization in three out of six. The number of primary care practices per capita had a significant effect on utilization in two ACGs. The supply of hospital outpatient services was significant in one ACG. CONCLUSIONS. Study findings reveal that some ACGs are price-sensitive and some are not. Policies aimed at changing levels of primary care use should taken into account whether varying cost-sharing will influence consumer behavior in the desired direction.     

Physician visits, hospitalizations, and socioeconomic status: ambulatory care sensitive conditions in a canadian setting

OBJECTIVE: To determine whether rates of physician visits for ambulatory care sensitive (ACS) conditions are lower for people of low-socioeconomic status than of high-socioeconomic status in an urban population with universal health care coverage. DATA SOURCES/STUDY SETTING: Physician claims and hospital discharge abstracts from fiscal years 1998 to 2001 for urban residents of Manitoba, Canada. The 1996 Canadian Census public use database provided neighborhood household income information. The study included all continuously enrolled urban residents in the Manitoba Health Services Insurance Plan. STUDY DESIGN: Twelve ACS conditions definable using 3-digit ICD-9-CM codes permitted cross-sectional and longitudinal comparison of ambulatory visits and hospitalizations. Neighborhood household income data provided a measure of socioeconomic status. DATA COLLECTION/EXTRACTION METHODS: Files were extracted from administrative data housed at the Manitoba Centre for Health Policy. PRINCIPAL FINDINGS: All conditions showed a socioeconomic gradient with residents of the lowest income neighborhoods having both more visits and more hospitalizations than their counterparts in higher income areas. Six of nine conditions with a sufficient N showed individuals living in the lowest income neighborhoods to have significantly more ambulatory visits before hospitalization for an ACS condition than did those in the most affluent neighborhoods. Many conditions showed a gradient in rate of hospitalization even after controlling for the number of ambulatory care visits. CONCLUSIONS: In the Canadian universal health care plan, the poor have reasonable access to ambulatory care for ACS conditions. Ambulatory care may be more effective in preventing hospitalizations among relatively affluent individuals than among the less well off.     

Differences in prevalence of pre-existing morbidity between injured and non-injured populations

OBJECTIVE: To identify and examine differences in pre-existing morbidity between injured and non-injured population-based cohorts. METHODS: Administrative health data from Manitoba, Canada, were used to select a population-based cohort of injured people and a sample of non-injured people matched on age, gender, aboriginal status and geographical location of residence at the date of injury. All individuals aged 18-64 years who had been hospitalized between 1988 and 1991 for injury (International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) code 800-995) (n = 21 032), were identified from the Manitoba discharge database. The matched non-injured comparison group comprised individuals randomly selected 1:1 from the Manitoba population registry. Morbidity data for the 12 months prior to the date of the injury were obtained by linking the two cohorts with all hospital discharge records and physician claims. RESULTS: Compared to the non-injured group, injured people had higher Charlson Comorbidity Index scores, 1.9 times higher rates of hospital admissions and 1.7 times higher rates of physician claims in the year prior to the injury. Injured people had a rate of admissions to hospital for a mental health disorder 9.3 times higher, and physician claims for a mental health disorder 3.5 times higher, than that of non-injured people. These differences were all statistically significant (P < 0.001). CONCLUSION: Injured people were shown to differ from the general non-injured population in terms of pre-existing morbidity. Existing population estimates of the attributable burden of injury that are obtained by extrapolating from observed outcomes in samples of injured cases may overestimate the magnitude of the problem.     

Ambulatory care groups and the profiling of primary care physician resource use: examining the application of case mix adjustments

A variety of profiling models and tools is utilized by payers, providers, and regulators to evaluate physician work, performance, and resource utilization. In physician profiling, the provider's pattern of practice is expressed as a rate of service or outcome. The article by Tucker, Weiner, Honigfeld, and Parton (this issue) compares the practice-based norms of primary care physicians by adjusting for case mix using ambulatory care groups (ACGs), a population-based classification method. Once the case mix is adjusted, the actual use of resources, as measured by overall charges, is compared with the expected value of resource use. In the Center for Research in Ambulatory Health Care Administration (CRAHCA) Physician Profiling Project, funded by The Robert Wood Johnson Foundation, physicians learn which services other physicians in their specialties perform. Physicians are able to compare their profiles with state and national level medians. The profiling project is one of the first demonstration projects in the field to profile ambulatory care practice patterns and collect patient demographics. An aspect of the project is to test the ACG classification system to data selected from 130 nonacademic practices representing over 5,000 physicians.     

Managing health services: how administrative data and population-based analyses can focus the agenda.

University-based researchers in Manitoba, Canada, have used administrative data routinely collected as part of the national health insurance plan to design an integrated database and population-based health information system. This information system is proving useful to policymakers for providing answers to such questions as: which populations need more physician services? Which need fewer? Are high-risk populations poorly served or do they have poor health outcomes despite being well served? Does high utilization represent overuse or utilization related to high need? More specifically, this system provides decision-makers with the capability to make critical comparisons across regions and subregions of residents' health status, socioeconomic risk characteristics, and use of hospitals, nursing homes, and physicians. The system permits analyses of demographic changes, expenditure patterns, and hospital performance in relation to the population served. The integrated database has also facilitated outcomes research across hospitals and counties, utilization review within a single hospital, and longitudinal research on health reform. A particularly interesting application to planning physician supply and distribution is discussed. The discussion highlights the strengths of integrated population-based information in analyzing the health care system and raising important questions about the relationship between health care and health.     

Conspicuous consumption: characterizing high users of physician services in one Canadian province

OBJECTIVES: To examine medical care use and costs, patterns of morbidity and co-morbidity, and other patient characteristics of high users of physician services in British Columbia. METHODS: This population-based study uses physician claims, hospital discharge summaries and vital statistics data linked at the level of the individual to compare characteristics of high users, other users and non-users of physician services in the Province of British Columbia, Canada. The study included all enrolled adults in the universal health care plan during fiscal year 1996/97. High users were defined as the most costly 5% of users of fee-reimbursed services. Key variables included age, sex, an ecological socio-economic status indicator and a comprehensive set of morbidity indicators, derived from the diagnoses recorded on the utilization records. RESULTS: The top 5% of users consumed a disproportionate 30% of spending on physician services. High users were overwhelmingly characterized by a significant burden of morbidity. Over 80% had at least six different types of morbidity during the study year compared with fewer than 20% of other users. High users were also much more likely to have major diagnoses that were both acute and chronic in nature. Co-morbidity involving psychosocial and chronic medical conditions was also very common. CONCLUSIONS: High users of physician services are overwhelmingly characterized by multiple and complex health problems. Policy tools based on a philosophy of deterrence such as cost-sharing are unlikely to have much impact on their costs and will likely do considerable harm.     

Self-rated health: a predictor of mortality among the elderly.

Data from the Manitoba Longitudinal Study on Aging (MLSA) were used to test the hypothesis that self-rated health (SRH) is a predictor of mortality independent of "objective health status" (OHS). Subjects were a random sample of non-institutionalized residents of Manitoba aged 65+ in 1971 (n = 3,128). A single item measure of SRH was obtained during a survey conducted in 1971; a baseline measure of OHS was derived from physician and self-reported conditions and health service utilization data. Occurrence and date of death during the years 1971-1977 were known. Analyses of the data revealed that, controlling for OHS, age, sex, life satisfaction, income and urban/rural residence, the risk of early mortality (1971-1973) and late mortality (1974-1977) for persons whose SRH was poor was 2.92 and 2.77 times that of those whose SRH was excellent. This increased risk of death associated with poor self-rated health was greater than that associated with poor OHS, poor life satisfaction, low income and being male. These findings provide empirical support for the long held, but inadequately substantiated, belief that the way a person views his health is importantly related to subsequent health outcomes.     

Evaluation of the equity of age-sex adjusted primary care capitation payments in Ontario, Canada

Objective

Several innovative primary care models have recently been introduced in Ontario, Canada. Two of these models are funded primarily through age-sex based capitation. There is concern that adjusting capitation rates for age and sex alone does not take into account the increased morbidity burden and health care needs that are associated with lower socioeconomic status. This study assesses the extent to which the current age-sex capitation rates in Ontario reflect health care needs of patients across socioeconomic status by comparing Ontario's age-sex adjusted capitation remuneration rate index with relative expected health care resource use by socioeconomic status (SES).

Methods

This study used administrative data collected by the Ontario Ministry of Health and Long-Term Care. The study sample was those patients who were enrolled to a FHN continuously from September 1, 2005 to August 31, 2006. Standardized expected health care utilization was calculated based on morbidity burden using The Johns Hopkins Adjusted Clinical Groups (ACG) Case-mix System and compared with standardized capitation rates across and within neighbourhood income quintiles.

Results

Among those in the lowest income group expected utilization was much higher than the age-sex capitation rates, while the opposite was true for those in the highest income group.

Conclusions

The findings suggests that under the physician reimbursement system used in Family Health Networks in Ontario, physicians are under-compensated for the health care needs of low income patients and over-compensated for the needs of high income patients. Adjusting capitation rates for morbidity burden in addition to age and sex may reduce incentives to preferentially enrol patients with higher socioeconomic status.     

Socioeconomic differences in use of medical care and antibiotics among schoolchildren in Sweden

BACKGROUND: Studies of mortality among children and adults in Sweden have demonstrated considerable socioeconomic differences. This paper describes socioeconomic patterns of physical morbidity and use of medical care and antibiotics in schoolchildren in Sweden. METHODS: A cross-sectional study based on parent interviews from the Swedish Survey of Living Conditions in 1996-1997 was used. The study population consisted of 3,557 children aged 6-15 years. RESULTS: Forty-five percent of the schoolchildren in the study were reported to have been absent from school because of illness at least once during the previous three months, 8% were taking regular medication and 10% had ever suffered from a chronic disorder. There were no indications of socioeconomic differences according to the education of the responding parent in morbidity or use of consultations with a physician. However, children in families where the responding parents had primary education only consumed antibiotics less often (OR 0.7 and CI: 0.5-0.9) when compared to children in families with post-secondary education. Children in rural areas used consultations with a physician less often and consumed less antibiotics (adjusted OR 0.7 and CI: 0.4-0.9 and 0.7 and CI: 0.5-0.9 respectively). CONCLUSION: No obvious patterns of socioeconomic inequality in physical morbidity or use of medical care were identified among schoolchildren in Sweden. Further studies are needed in order to explain the social inequality in consumption of antibiotics among schoolchildren in Sweden and to describe social and regional patterns of psychiatric, behavioural and psychosomatic morbidity.     

Variations in health services utilization by primary care patients.

This study analyses how both ambulatory care groups (ACGs) and physician characteristics explain the variability in health-service use among primary care patients in Spain. During the period 1996-1997, data derived from 52,152 patients and their 38 respective primary care physicians were collected. The response variables were as follows: number of visits; diagnostic tests requested; and referrals to a specialist. ACGs are an important variable that should be taken into account in order to explain health-service utilization. As for professionals, age and the post they hold are essential factors. Most of the unexplained variability is caused by patient characteristics.     

The influence of environmental and health indicators on premature mortality: An empirical analysis of the City of Toronto's 140 neighborhoods

The objective of this paper was to assess the link between premature mortality and a combination of neighbourhood contextual (environmental and health) and compositional (socioeconomic and demographic) characteristics. We statistically and spatially examined six environmental variables (ultrafine particles, carcinogenic and non-carcinogenic pollutants, pollution released to air, tree cover, and walkability index), six health service indicators (number health providers, breast, colorectal and cervical cancer screening uptake rates, student nutrition program uptake rates, and healthy food index), and eight socioeconomic indicators (total income, Gini coefficient, two age categories – below and above 40 years, proportion of females to males, visible minorities, Indigenous peoples, education, less than grade 9) among 140 neighbourhoods of the City of Toronto in Ontario (Canada). We applied principal component analysis to identify patterns and to reduce the number of explanatory variables into combined component axes that represent unique variation in these confounded and overlapping factors. We then applied regression analysis to model the relationship between the indices of enviro-health and socioeconomics and their potential relationship with premature mortality. Residual spatial analysis was used to investigate any remaining spatial structure (such as neighbourhoods with higher residual premature mortality rates). Neighbourhood Equity Index was correlated with our enviro-health and socioeconomic indices. Premature mortality within neighbourhoods was predicted by poor cancer screenings, pollution, lack of tree canopy, increased uptake of student nutrition programs and high walkability index. A negative association between premature mortality and pollution was associated low walkability index and presence of visible minorities within neighbourhoods. There was some unexplained residual spatial variation in our model of premature mortality - especially along the shores of Lake Ontario and in neighbourhoods with major highways or road corridors: premature mortality in Toronto neighbourhoods was higher than expected along highway-corridor neighbourhoods and shorelines. Our analysis revealed a significant relationship between neighbourhood contextual features – both environmental and health – and premature mortality, suggesting that these contextual components of neighbourhoods can predict rates of urban premature mortality in Toronto.     

Comparing self-reported health status and diagnosis-based risk adjustment to predict 1- and 2 to 5-year mortality

OBJECTIVES: To compare the ability of two diagnosis-based risk adjustment systems and health self-report to predict short- and long-term mortality. DATA SOURCES/STUDY SETTING: Data were obtained from the Department of Veterans Affairs (VA) administrative databases. The study population was 78,164 VA beneficiaries at eight medical centers during fiscal year (FY) 1998, 35,337 of whom completed an 36-Item Short Form Health Survey for veterans (SF-36V) survey. STUDY DESIGN: We tested the ability of Diagnostic Cost Groups (DCGs), Adjusted Clinical Groups (ACGs), SF-36V Physical Component score (PCS) and Mental Component Score (MCS), and eight SF-36V scales to predict 1- and 2-5 year all-cause mortality. The additional predictive value of adding PCS and MCS to ACGs and DCGs was also evaluated. Logistic regression models were compared using Akaike's information criterion, the c-statistic, and the Hosmer-Lemeshow test. PRINCIPAL FINDINGS: The c-statistics for the eight scales combined with age and gender were 0.766 for 1-year mortality and 0.771 for 2-5-year mortality. For DCGs with age and gender the c-statistics for 1- and 2-5-year mortality were 0.778 and 0.771, respectively. Adding PCS and MCS to the DCG model increased the c-statistics to 0.798 for 1-year and 0.784 for 2-5-year mortality. CONCLUSIONS: The DCG model showed slightly better performance than the eight-scale model in predicting 1-year mortality, but the two models showed similar performance for 2-5-year mortality. Health self-report may add health risk information in addition to age, gender, and diagnosis for predicting longer-term mortality.     

Wide range of socioeconomic factors associated with mortality among cities in Japan

The aim of this study was to identify socioeconomic factors associated with mortality among cities in Japan. Sex-specific and age-adjusted mortality rates for 1990 and 1995 were calculated by 779 local administrative units across the nation. One hundred indicators related to socioeconomic factors were compiled and divided into eight categories: economy, education, living conditions, vegetation and open space, transport, preventive activities, medical care and demography. Composite socioeconomic indices were formulated using factor analysis of the socioeconomic indicators by category, and the association between the indices and mortality rates was examined by correlation analysis and multiple regression analysis. Nineteen composite socioeconomic indices were obtained from factor analysis, and all indices except educational expenditure-related index were significantly correlated with mortality rates. Unemployment, old housing, primary health resources and density were independently positively associated, and higher education, public library activity, health check-up participation and population growth were independently negatively associated with both 1990 and 1995 male mortality rates. For female mortality, higher income, unemployment, spacious dwelling, old housing, less vegetation, road facility, numbers of cars per population, primary health resources and density were independently positively associated, and higher education, public library activity and health check-up participation were independently negatively associated. The relationship between mortality and socioeconomic conditions was stronger in males than in females, and higher income and less vegetation were associated with higher mortality only for females. The present study demonstrated a close link between mortality and a wide range of socioeconomic conditions by using a number of indicators compiled from various data sources. The results promote a deeper understanding of socioeconomic health determinants and development of multi-sectoral health policy to improve population health.     

Common comorbidity scales were similar in their ability to predict health care costs and mortality

OBJECTIVE: To compare the ability of commonly used measures of medical comorbidity (ambulatory care groups [ACGs], Charlson comorbidity index, chronic disease score, number of prescribed medications, and number of chronic diseases) to predict mortality and health care costs over 1 year. STUDY DESIGN AND SETTING: A prospective cohort study of community-dwelling older adults (n=3,496) attending a large primary care practice. RESULTS: For predicting health care charges, the number of medications had the highest predictive validity (R(2)=13.6%) after adjusting for demographics. ACGs (R(2)=16.4%) and the number of medications (15.0%) had the highest predictive validity for predicting ambulatory visits. ACGs and the Charlson comorbidity index (area under the receiver operator characteristic [ROC] curve=0.695-0.767) performed better than medication-based measures (area under the ROC curve=0.662-0.679) for predicting mortality. There is relatively little difference, however, in the predictive validity across these scales. CONCLUSION: In an outpatient setting, a simple count of medications may be the most efficient comorbidity measure for predicting utilization and health-care charges over the ensuing year. In contrast, diagnosis-based measures have greater predictive validity for 1-year mortality. Current comorbidity measures, however, have only poor to moderate predictive validity for costs or mortality over 1 year.     

Ten-year health service use outcomes in a population-based cohort of 21,000 injured adults: the Manitoba injury outcome study

OBJECTIVE: To quantify long-term health service use (HSU) following non-fatal injury in adults. METHODS: A retrospective, population-based, matched cohort study identified an inception cohort (1988-91) of injured people who had been hospitalized (ICD-9-CM 800-995) aged 18-64 years (n = 21 032) and a matched non-injured comparison group (n = 21 032) from linked administrative data from Manitoba, Canada. HSU data (on hospitalizations, cumulative length of stay, physician claims and placements in extended care services) were obtained for the 12 months before and 10 years after the injury. Negative binomial and Poisson regressions were used to quantify associations between injury and long-term HSU. FINDINGS: Statistically significant differences in the rates of HSU existed between the injured and non-injured cohorts for the pre-injury year and every year of the follow-up period. After controlling for pre-injury HSU, the attributable risk percentage indicated that 38.7% of all post-injury hospitalizations (n = 25 183), 68.9% of all years spent in hospital (n = 1031), 21.9% of physician claims (n = 269 318) and 77.1% of the care home placements (n = 189) in the injured cohort could be attributed to being injured. CONCLUSION: Many people who survive the initial period following injury, face long periods of inpatient care (and frequent readmissions), high levels of contact with physicians and an increased risk of premature placement in institutional care. Population estimates of the burden of injury could be refined by including long-term non-fatal health consequences and controlling for the effect of pre-injury comorbidity.     

Health care utilization in the years prior to death

The impact of an aging population on the health care system is a primary speculative concern for health policy. Unique data from a large sample of 4,263 decedents aged 45 years and over in Manitoba, Canada, describe actual utilization in the four years prior to death: all hospitalizations, nursing home stays, and ambulatory physician contacts. Total expenditures associated with dying do increase with age, but even among the very elderly many deaths have few expenditure consequences. Apocalyptic scenarios for the health care system may be premature.     

Can the Quality of Care in Family Practice Be Measured Using Administrative Data?

OBJECTIVE: To explore the feasibility of using administrative data to develop process indicators for measuring quality in primary care. DATA SOURCES/STUDY SETTING: The Population Health Research Data Repository (Repository) housed at the Manitoba Centre for Health Policy which includes physician claims, hospital discharge abstracts, pharmaceutical use (Drug Program Information Network (DPIN)), and the Manitoba Immunization Monitoring Program (MIMS) for all residents of Manitoba, Canada who used the health care system during the 2001/02 fiscal year. Family physicians were identified from the Physician Resource Database. Indicators were developed based on a literature review and focus group validation. DATA COLLECTION/EXTRACTION METHODS: Data files were extracted from administrative data available in the Repository. We extracted data based on the ICD-9-CM codes and ATC-class drugs prescribed and then linked them to the Physician Resource Database. Physician practices were defined by allocating patients to their most responsible physician. Every family physician in Manitoba that met the inclusion criteria (having either 5 or 10 eligible patients depending on the indicator) was 'scored' on each indicator. Physicians were then grouped according to the proportion of the patients allocated to their practice who received the recommended care for the specific indicator. PRINCIPAL FINDINGS: Using administrative health data we were able to develop and measure eight indicators of quality of care covering both preventive care services and chronic disease management. The number of eligible physicians and patients varied for each indicator as did the percent of patients with recommended care, per physician. For example, the childhood immunization indicator included 544 physicians who, on average, provided immunization for 65 percent of their patients. CONCLUSIONS: Quality of care provided by family physicians can be measured using administrative data. Despite the limitations addressed in this paper, this work establishes a practical methodology to measure quality of care provided by family physicians that can be used for quality improvement initiatives.