Mapping the prevalence of cancer risk factors at the small area level in Australia

BACKGROUND

Cancer is a significant health issue globally and it is well known that cancer risk varies geographically. However in many countries there are no small area-level data on cancer risk factors with high resolution and complete reach, which hinders the development of targeted prevention strategies.

METHODS

Using Australia as a case study, the 2017-2018 National Health Survey was used to generate prevalence estimates for 2221 small areas across Australia for eight cancer risk factor measures covering smoking, alcohol, physical activity, diet and weight. Utilising a recently developed Bayesian two-stage small area estimation methodology, the model incorporated survey-only covariates, spatial smoothing and hierarchical modelling techniques, along with a vast array of small area-level auxiliary data, including census, remoteness, and socioeconomic data. The models borrowed strength from previously published cancer risk estimates provided by the Social Health Atlases of Australia. Estimates were internally and externally validated.

RESULTS

We illustrated that in 2017-2018 health behaviours across Australia exhibited more spatial disparities than previously realised by improving the reach and resolution of formerly published cancer risk factors. The derived estimates revealed higher prevalence of unhealthy behaviours in more remote areas, and areas of lower socioeconomic status; a trend that aligned well with previous work.

CONCLUSIONS

Our study addresses the gaps in small area level cancer risk factor estimates in Australia. The new estimates provide improved spatial resolution and reach and will enable more targeted cancer prevention strategies at the small area level. Furthermore, by including the results in the next release of the Australian Cancer Atlas, which currently provides small area level estimates of cancer incidence and relative survival, this work will help to provide a more comprehensive picture of cancer in Australia by supporting policy makers, researchers, and the general public in understanding the spatial distribution of cancer risk factors. The methodology applied in this work is generalisable to other small area estimation applications and has been shown to perform well when the survey data are sparse.

Geographical variation in compulsory hospitalisation - ethical challenges

BACKGROUND

Compulsory hospitalisation in mental health care restricts patients' liberty and is experienced as harmful by many. Such hospitalisations continue to be used due to their assumed benefit, despite limited scientific evidence. Observed geographical variation in compulsory hospitalisation raises concern that rates are higher and lower than necessary in some areas.

METHODS/DISCUSSION

We present a specific normative ethical analysis of how geographical variation in compulsory hospitalisation challenges four core principles of health care ethics. We then consider the theoretical possibility of a "right", or appropriate, level of compulsory hospitalisation, as a general norm for assessing the moral divergence, i.e., too little, or too much. Finally, we discuss implications of our analysis and how they can inform the future direction of mental health services.

Small Area Analysis of Veterans Affairs Mental Health Services Data

OBJECTIVE

To identify geographic variation in mental health service use in the Department of Veterans Affairs (VA), the authors constructed utilization-based VA mental health service areas (MHSAs) for outpatient treatment and mental health referral regions (MHRRs) for residential and acute inpatient treatment.

METHODS

MHSAs are empirically derived geographic groupings of one or more counties containing one or more VA outpatient mental health clinics. For each county within an MHSA, patients received most of their VA-provided outpatient mental health care within that MHSA. MHSAs were aggregated into MHRRs according to where VA users in each MHSA received most of their residential and acute inpatient mental health care. Attribution loyalty was evaluated with the localization index-the fraction of VA users living in each geographic area who used their designated MHSA and MHRR facility. Variation in outpatient mental health visits and in acute inpatient and residential mental health stays was determined for the 2008-2018 period.

RESULTS

A total of 441 MHSAs were aggregated to 115 MHRRs (representing 3,909,080 patients with 52,372,303 outpatient mental health visits). The mean±SD localization index was 59.3%±16.4% for MHSAs and 67.8%±12.7% for MHRRs. Adjusted outpatient mental health visits varied from a mean of 0.88 per year in the lowest quintile of MHSAs to 3.14 in the highest. Combined residential and acute inpatient days varied from 0.29 to 1.79 between the lowest and highest quintiles.

CONCLUSIONS

MHSAs and MHRRs validly represented mental health utilization patterns in the VA and displayed considerable variation in mental health service provision across different locations.

A comparison of two neighborhood-level socioeconomic indexes in the United States

socioeconomic indexes that capture information about wealth, education, employment, and housing are in wide use in public health. Here we compare the widely used Area Deprivation Index (ADI) to the Yost index. Though they are derived largely from the same data, there are substantial differences between the two. Examination of the geographic areas where the two indexes are most dissimilar suggest that the Yost index has greater face validity and that the ADI is highly sensitive to locations with incomplete census data and with census data containing outliers.

Wide systematic variations in potentially avoidable hospitalisations of chronically ill patients: Ecological study of basic health areas and healthcare areas

BACKGROUND AND OBJECTIVE

Potentially avoidable hospitalisations (PAHs) due to chronic conditions are a healthcare problem that could reflect healthcare of insufficient quality. This study reports the systematic variations in PAHs for the collection of providers of the Spanish National Health System.

MATERIALS AND METHODS

We conducted an ecological study on government data, analysing the systematic variation in PAHs for 6 chronic conditions during 2013-2015. To determine the variation, we performed a small area analysis using Bayesian methodology.

RESULTS

Between 2013 and 2015, 439,878 admissions for PAHs were recorded in the Spanish National Health System. There was an up to 4-fold difference in PAH rates between certain basic health areas (BHA), with highly variable differences depending on the analysed condition. Forty percent of the BHAs showed a greater than expected risk of PAH. Beyond the systematic variation observed between BHAs, the healthcare areas of the patients' residence explained 33% of the variation in PAHs. We observed specific differences in these general results according to clinical condition, age and sex.

CONCLUSIONS

The wide systematic variation in PAHs suggests a problem of quality in the care provided to chronically ill patients by the providers of healthcare areas in Spain. Identifying and analysing these areas and other healthcare areas with better results could provide a reference for improving the care of other suppliers with poorer performance.

Hospital, health, and community burden after oil refinery fires, Richmond, California 2007 and 2012

INTRODUCTION

Emergency Departments experience a significant census burst after disasters. The aim of this study is to describe patient presentations at Emergency Departments in Contra Costa County, California following chemical release incidents at an oil refinery in 2007 and 2012. Specific areas of focus include hospital and community burden with an emphasis on disease classes.

METHODS

Searching 4 weeks before through 4 weeks after each event, Emergency Department abstracts identified patients living in Contra Costa County and seeking care there or in neighboring Alameda County. City and ZIP-code of residence established proximity to the refinery. This provided the following contrast groups: Event (2007, 2012), time (before, after), location (bayside, rest of county), and within bayside, warned or not warned to shelter in place. Using the Multi-Level Clinical Classification Software, we classified primary health groups recorded 4 weeks before and after the events, then summarized the data, calculated rates, and made tables, graphs, and maps to highlight findings.

RESULTS

Number of visits meeting selection criteria totalled 105020 records. Visits increased modestly but statistically significantly after the 2007 incident. After the 2012 incident, two Emergency Departments took the brunt of the surge. Censuses increased from less than 600 a week each to respectively 5719 and 3072 the first week, with the greatest number 2 days post-event. It took 4 weeks for censuses to return to normal. The most common diagnosis groups that spiked were nervous/sensory, respiratory, circulatory, and injury. Bayside communities had statistically significant increases in residents seeking care. Specifically, visits of residents in warned communities nearest the refinery increased by a factor of 3.7 while visits of residents in other nearby un-warned communities increased by a factor of 1.5.

CONCLUSIONS

The 2012 Emergency Department census peaked in the first week and did not return to normal for 4 weeks. Diagnoses changed to reflect conditions associated with reactions to chemical exposures. Surrounding communities and nearby hospitals experienced significant emergent burdens. In addition to changes from such events in patient diagnoses and community burden, the discussion highlights the long-term implications of failures to require adequate monitoring and warning systems and failures of health planning.

Geographic variation within the military health system

Background

This study seeks to quantify variation in healthcare utilization and per capita costs using system-defined geographic regions based on enrollee residence within the Military Health System (MHS).

Methods

Data for fiscal years 2007 – 2010 were obtained from the Military Health System under a data sharing agreement with the Defense Health Agency (DHA). DHA manages all aspects of the Department of Defense Military Health System, including TRICARE. Adjusted rates were calculated for per capita costs and for two procedures with high interest to the MHS- back surgery and Cesarean sections for TRICARE Prime and Plus enrollees. Coefficients of variation (CoV) and interquartile ranges (IQR) were calculated and analyzed using residence catchment area as the geographic unit. Catchment areas anchored by a Military Treatment Facility (MTF) were compared to catchment areas not anchored by a MTF.

Results

Variation, as measured by CoV, was 0.37 for back surgery and 0.13 for C-sections in FY 2010- comparable to rates documented in other healthcare systems. The 2010 CoV (and average cost) for per capita costs was 0.26 ($3,479.51). Procedure rates were generally lower and CoVs higher in regions anchored by a MTF compared with regions not anchored by a MTF, based on both system-wide comparisons and comparisons of neighboring areas.

Conclusions

In spite of its centrally managed system and relatively healthy beneficiaries with very robust health benefits, the MHS is not immune to unexplained variation in utilization and cost of healthcare.

Dying among older adults in Switzerland: who dies in hospital, who dies in a nursing home?

BACKGROUND

Institutional deaths (hospitals and nursing homes) are an important issue because they are often at odds with patient preference and associated with high healthcare costs. The aim of this study was to examine deaths in institutions and the role of individual, regional, and healthcare supply characteristics in explaining variation across Swiss Hospital Service Areas (HSAs).

METHODS

Retrospective study of individuals ≥66 years old who died in a Swiss institution (hospital or nursing homes) in 2010. Using a two-level logistic regression analysis we examined the amount of variation across HSAs adjusting for individual, regional and healthcare supply measures. The outcome was place of death, defined as death in hospital or nursing homes.

RESULTS

In 2010, 41,275 individuals ≥66 years old died in a Swiss institution; 54 % in nursing homes and 46 % in hospitals. The probability of dying in hospital decreased with increasing age. The OR was 0.07 (95 % CI: 0.05-0.07) for age 91+ years compared to those 66-70 years. Living in peri-urban areas (OR = 1.06 95 % CI: 1.00-1.11) and French speaking region (OR = 1.43 95 % CI: 1.22-1.65) was associated with higher probability of hospital death. Females had lower probability of death in hospital (OR = 0.54 95 % CI: 0.51-0.56). The density of ambulatory care physicians (OR = 0.81 95 % CI: 0.67-0.97) and nursing homes beds (OR = 0.67 95 % CI: 0.56-0.79) was negatively associated with hospital death. The proportion of dying in hospital varied from 38 % in HSAs with lowest proportion of hospital deaths to 60 % in HSAs with highest proportion of hospital deaths (1.6-fold variation).

CONCLUSIONS

We found evidence for variation across regions in Switzerland in dying in hospital versus nursing homes, indicating possible overuse and underuse of end of life (EOL) services.

Examining regional variation in health care spending in British Columbia, Canada

Examining regional variation in health care spending may reveal opportunities for improved efficiency. Previous research has found that health care spending and service use vary substantially from place to place, and this is often not explained by differences in the health status of populations or by better outcomes in higher-spending regions, but rather by differences in intensity of service provision. Much of this research comes from the United States. Whether similar patterns are observed in other high-income countries is not clear. We use administrative data on health care use, covering the entire population of the Canadian province of British Columbia, to examine how and why health care spending varies among health regions. Pricing and insurance coverage are constant across the population, and we adjust for patient-level age, sex, and recorded diagnoses. Without adjusting for differences in population characteristics, per-capita spending is 50% higher in the highest-spending region than in the lowest. Adjusting for population characteristics as well as the very different environments for health service delivery that exist among metropolitan, non-metropolitan, and remote regions of the province, this falls to 20%. Despite modest variation in total spending, there are marked differences in mortality. In this context, it appears that policy reforms aimed at system-wide quality and efficiency improvement, rather than targeted at high-spending regions, will likely prove most promising.

A systematic review of medical practice variation in OECD countries

BACKGROUND

Major variations in medical practice have been documented internationally. Variations raise questions about the quality, equity, and efficiency of resource allocation and use, and have important implications for health care and health policy.

OBJECTIVE

To perform a systematic review of the peer-reviewed literature on medical practice variations in OECD countries.

METHODS

We searched MEDLINE to find publications on medical practice variations in OECD countries published between 2000 and 2011. We present an overview of the characteristics of published studies as well as the magnitude of variations for select high impact conditions.

RESULTS

A total of 836 studies were included. Consistent with the gray literature, there were large variations across regions, hospitals and physician practices for almost every condition and procedure studied. Many studies focused on high-impact conditions, but very few looked at the causes or outcomes of medical practice variations.

CONCLUSION

While there were an overwhelming number of publications on medical practice variations the coverage was broad and not often based on a theoretical construct. Future studies should focus on conditions and procedures that are clinically important, policy relevant, resource intensive, and have high levels of public awareness. Further study of the causes and consequences of variations is important.

Health, policy and geography: insights from a multi-level modelling approach

Improving the health and wellbeing of citizens ranks highly on the agenda of most governments. Policy action to enhance health and wellbeing can be targeted at a range of geographical levels and in England the focus has tended to shift away from the national level to smaller areas, such as communities and neighbourhoods. Our focus is to identify the potential for targeting policy interventions at the most appropriate geographical levels in order to enhance health and wellbeing. The rationale is that where variations in health and wellbeing indicators are larger, there may be greater potential for policy intervention targeted at that geographical level to have an impact on the outcomes of interest, compared with a strategy of targeting policy at those levels where relative variations are smaller. We use a multi-level regression approach to identify the degree of variation that exists in a set of health indicators at each level, taking account of the geographical hierarchical organisation of public sector organisations. We find that for each indicator, the proportion of total residual variance is greatest at smaller geographical areas. We also explore the variations in health indicators within a hierarchical level, but across the geographical areas for which public sector organisations are responsible. We show that it is feasible to identify a sub-set of organisations for which unexplained variation in health indicators is significantly greater relative to their counterparts. We demonstrate that adopting a geographical perspective to analyse the variation in indicators of health at different levels offers a potentially powerful analytical tool to signal where public sector organisations, faced increasingly with many competing demands, should target their policy efforts. This is relevant not only to the English context but also to other countries where responsibilities for health and wellbeing are being devolved to localities and communities.

•

Policies to enhance wellbeing may be targeted at different hierarchical levels such as individual, neighbourhood or region.

•

Organisations responsible for implementing such policies also operate in geographically defined administrative boundaries.

•

We explore the variation in a range of health-related indicators at a number of geographical hierarchical levels.

•

We find evidence of variation between organisations within the same hierarchy and also between hierarchical levels.

•

The results direct policymakers in public sector organisations to target their policy efforts where impact may be greatest.

Implications and limitations of appropriateness studies for chiropractic

OBJECTIVE

The appropriate role for chiropractic in US health care has not been established, but third-party payors and public policy makers must make decisions about the appropriate role for chiropractors in health care systems and for the services that chiropractors provide. Appropriateness studies for chiropractic may inform those decisions. The purpose of this article is to discuss the implications and limitations of appropriateness studies for chiropractic.

DISCUSSION

We reviewed the general context for assessment of the appropriateness and the application of appropriateness studies to chiropractic in particular. We evaluated the implications and limitations for chiropractic of methods of small area analysis and the RAND-UCLA Appropriateness Method. The RAND-UCLA Appropriateness Method has been applied to the evaluation of spinal manipulation. Regional variations in chiropractic utilization have yet to be described through small area analysis, but these methods appear to hold some potential for assessing the appropriateness of chiropractic care. Both small area analysis and the RAND-UCLA method offer limited possibilities for the assessment of chiropractic appropriateness.

CONCLUSION

Future assessment of the appropriate role for chiropractic in US health care will raise issues beyond the scope of previous appropriateness studies. Studying the appropriate role for chiropractic will require consideration of the clinical discipline in its entirety, rather than individual consideration of specific interventions. A fair assessment of chiropractic appropriateness will require new evidence and perhaps new research methodologies.