Small Area Variation Analysis Research Articles

공간분석을 활용한 식 행동 및 관련요인을 이용한 예측분석: 2016 지역사회건강조사자료

This study was carried out to investigate the magnitude and pattern of small area variations as defined by Si-gun-gu and to predict factors related to eating behavior and chronic disease using the 2016 Community Health Survey (CHS) data. The subjects of the analysis were the population of 254 counties surveyed in the 2009 CHS. The magnitude and pattern of area variations in the given eating behavior and chronic diseases were examined using Extremal Quotient (EQ) and Coefficients of Variation (CV) of Small Area Variation Analysis (SAVA). A spatial autocorrelation regression model was used to examine the related factors with these variations. Assessment of breakfast eating frequency in each community indicated that the lowest eating frequency per week was 4.624 while the highest eating frequency was 6.739. People in urban areas had a lower number of breakfast eating days than those in rural areas. When a spatial analysis of general characteristics and breakfast eating frequency was conducted, it was predicted that the people in the areas with a higher female population, age, and income level would have a higher number of breakfast eating days. As for the correlation between chronic disease diagnosis and breakfast eating days, it was predicted that those who were diagnosed with diabetes would eat breakfast. The analysis of general characteristics and usual salt consumption levels predicted that the people in areas with a higher male population, age, and income level would consume less salt. As for the correlation between chronic diseases and salt consumption levels, it was predicted that those who were never diagnosed with hypertension would take less salt. These results will help us in the development of policies for population-based health promotion through a reduction of the gap in eating behavior indices between areas. Further research is needed to build accurate and reliable models of CHS data.

Degree of regional variation and effects of health insurance-related factors on the utilization of 24 diverse healthcare services - a cross-sectional study

BackgroundRegional variation in healthcare utilization could reflect unequal access to care, which may lead to detrimental consequences to quality of care and costs. The aims of this study were to a) describe the degree of regional variation in utilization of 24 diverse healthcare services in eligible populations in Switzerland, and b) identify potential drivers, especially health insurance-related factors, and explore the consistency of their effects across the services.MethodsWe conducted a cross-sectional study using health insurance claims data for the year of 2014. The studied 24 healthcare services were predominantly outpatient services, ranging from screening to secondary prevention. For each service, a target population was identified based on applicable clinical recommendations, and outcome variable was the use of the service. Possible influencing factors included patients’ socio-demographics, health insurance-related and clinical characteristics. For each service, we performed a comprehensive methodological approach including small area variation analysis, spatial autocorrelation analysis, and multilevel multivariable modelling using 106 mobilité spaciale regions as the higher level. We further calculated the median odds ratio in model residuals to assess the unexplained regional variation.ResultsUnadjusted utilization rates varied considerably across the 24 healthcare services, ranging from 3.5% (osteoporosis screening) to 76.1% (recommended thyroid disease screening sequence). The effects of health insurance-related characteristics were mostly consistent. A higher annual deductible level was mostly associated with lower utilization. Supplementary insurance, supplementary hospital insurance and having chosen a managed care model were associated with higher utilization of most services. Managed care models showed a tendency towards more recommended care. After adjusting for multiple influencing factors, the unexplained regional variation was generally small across the 24 services, with all MORs below 1.5.ConclusionsThe observed utilization rates seemed suboptimal for many of the selected services. For all of them, the unexplained regional variation was relatively small. Our findings confirmed the importance and consistency of effects of health insurance-related factors, indicating that healthcare utilization might be further optimized through adjustment of insurance scheme designs. Our comprehensive approach aids in the identification of regional variation and influencing factors of healthcare services use in Switzerland as well as comparable settings worldwide.

Geographic variation in the provision of naloxone by pharmacies in Ontario, Canada: A population-based small area variation analysis

BackgroundRegional variation in pharmacy-dispensed naloxone rates could create access disparities that undermine the effectiveness of this approach. We explored individual and public health unit (PHU)-level determinants of regional variation in naloxone distribution through the Ontario Naloxone Program for Pharmacies. MethodsWe conducted a population-based study between April 1, 2017 and March 31, 2018. We calculated age- and sex-standardized pharmacy-dispensed naloxone rates for the 35 Ontario PHUs, and identified determinants of these rates using generalized estimating equations negative binomial regression. ResultsThe age- and sex-standardized pharmacy-dispensed naloxone rate in Ontario was 5.5 (range 1.8–11.6) kits per 1000 population. Variables associated with higher naloxone dispensing rates included opioid use disorder history [rate ratio (RR) 2.27; 95% confidence interval (CI) 1.75–2.96], opioid agonist therapy (RR 11.17; 95% CI 7.15–17.44), and PHU opioid overdose rate (RR 1.09 per 10 deaths; 95% CI 1.06–1.13). Pharmacy-dispensed naloxone rates were lower in rural areas (RR 0.83; 95% CI 0.73–0.94) and among individuals dispensed one (RR 0.72; 95% CI 0.65–0.79), two to five (RR 0.67; 95% CI 0.54–0.84) or 6–10 (RR 0.92; 95% CI 0.74–1.14) opioids in the prior year relative to those receiving no opioids. ConclusionPharmacy-dispensed naloxone programs are important components of a public health response to the opioid overdose crisis. We found considerable variation in pharmacy-dispensed naloxone rates that could limit program effectiveness, particularly in rural settings with limited access to health and harm reduction services..

Unwarranted regional variation in vertebroplasty and kyphoplasty in Switzerland: A population-based small area variation analysis.

BackgroundPercutaneous vertebroplasty (VP) and balloon kyphoplasty (BKP) for treating painful osteoporotic vertebral fractures are controversial.ObjectiveWe assessed the regional variation in the use of VP/BKP in Switzerland.MethodsWe conducted a population-based small area variation analysis using patient discharge data for VP/BKP from all Swiss hospitals and Swiss census data for calendar years 2012/13. We derived hospital service areas (HSAs) by analyzing patient flows, assigning regions from which most residents were discharged to the same VP/BKP specific HSA. We calculated age-/sex-standardized mean VP/BKP-rates and measures of regional variation (extremal quotient [EQ], systematic component of variation [SCV]). We estimated the reduction in variation of VP/BKP rates using negative binomial regression, with adjustment for patient demographic and regional socioeconomic factors (socioeconomic status, urbanization, and language region). We considered the residual, unexplained variation most likely to be unwarranted.ResultsOverall, 4955 VP/BKPs were performed in Switzerland in 2012/13. The age-/sex-standardized mean VP/BKP rate was 4.6/10,000 persons and ranged from 1.0 to 10.1 across 26 HSAs. The EQ was 10.2 and the SCV 57.6, indicating a large variation across VP/BKP specific HSAs. After adjustment for demographic and socioeconomic factors, the total reduction in variance was 32.2% only, with the larger part of the variation remaining unexplained.ConclusionsWe found a 10-fold variation in VP/BKP rates across Swiss VP/BKP specific HSAs. As only one third of the variation was explained by differences in patient demographics and regional socioeconomic factors, VP/BKP in the highest-use areas may, at least partially, represent overtreatment.

Variations in osteoporosis medication utilization. A population-based ecological cross-sectional study in the region of Valencia, Spain

Little is known about the contextual variability in osteoporosis medication utilization. Our aims were 1) to describe variations in utilization and spending on osteoporotic medication between the Primary Care Health Zones (PHZ) of the Valencia region, Spain, 2) to analyze observed variations using Small Area Variation Analysis methods, and 3) to quantify the influence of the specialized care level on variations in utilization. We conducted a population-based cross-sectional ecological study of expenditure and utilization of five therapeutic groups marketed as osteoporosis treatments in Spain in 2009. The unit of analysis was the PHZ (in total 240) nested in the 23 Hospital Healthcare Departments (HHD) of the region of Valencia, covering a population of about 4.9 million inhabitants. Drug utilization was measured by dispensed Defined Daily Dose per 1000 women aged 50 years old and over and day (DID) per PHZ and cost was measured by the annual osteoporosis drug cost per woman aged 50 and older as well as the average price of DDD (Defined Daily Dose) in each PHZ. We calculated Indirect Standardized Drug Utilization Ratios (ISR) and we used Spearman’s correlation to analyze associations between the ISRs of the different therapies. The average osteoporosis drug consumption was 119.1 DID, ranging from 77.6 to 171.3 DID (2.2 times higher) between PHZs in the 5th and 95th percentiles. Annual expenditure also showed a two-fold variation among PHZs. Average prices of the DDD by therapeutic group showed very low or no variation, although they differed substantially among therapeutic groups. Regarding the standardized consumption of osteoporotic drugs, HHDs explained a substantial part (39%) of the variance among PHZs. In conclusion, there is considerable variability in the volume and choice of anti-osteoporotic treatments between PHZs. with HHDs explaining an important proportion of the variation in utilization. Interventions aimed at reducing variation to improve appropriate care should take into account both the PHZ and HHD levels of care.

Variation in Antiosteoporotic Drug Prescribing and Spending Across Spain. A Population-Based Ecological Cross-Sectional Study.

Introduction: Evidence has shown that utilization of antiosteoporotic medications does not correspond with risk, and studies on other therapies have shown that adequacy of pharmaceutical prescribing might vary between regions. Nevertheless, very few studies have addressed the variability in osteoporotic drug consumption. We aimed to describe variations in pharmaceutical utilization and spending on osteoporotic drugs between Health Areas (HA) in Spain.Methods: Population-based cross-sectional ecological study of expenditure and utilization of the five therapeutic groups marketed for osteoporosis treatment in Spain in 2009. Small area variation analysis (SAVA) methods were used. The units of analysis were the 168 HA of 13 Spanish regions, including 7.2 million women aged 50 years and older. The main outcomes were the defined daily dose (DDD) per 1000 inhabitants and day (DDD/1000/Day) dispensed according to the pharmaceutical claims reimbursed, and the expenditure on antiosteoporotics at retail price per woman ≥50 years old and per year.Results: The average osteoporosis drug consumption was 116.8 DDD/1000W/Day, ranging from 78.5 to 158.7 DDD/1000W/Day between the HAs in the 5th and 95th percentiles. Seventy-five percent of the antiosteoporotics consumed was bisphosphonates, followed by raloxifene, strontium ranelate, calcitonins, and parathyroid hormones including teriparatide. Regarding variability by therapeutic groups, biphosphonates showed the lowest variation, while calcitonins and parathyroid hormones showed the highest variation. The annual expenditure on antiosteoporotics was €426.5 million, translating into an expenditure of €59.2 for each woman ≥50 years old and varying between €38.1 and €83.3 between HAs in the 5th and 95th percentiles. Biphosphonates, despite accounting for 79% of utilization, only represented 63% of total expenditure, while parathyroid hormones with only 1.6% of utilization accounted for 15% of the pharmaceutical spending.Conclusion: This study highlights a marked geographical variation in the prescription of antiosteoporotics, being more pronounced in the case of costly drugs such as parathyroid hormones. The differences in rates of prescribing explained almost all of the variance in drug spending, suggesting that the difference in prescription volume between territories, and not the price of the drugs, is the main source of variation in this setting. Data on geographical variation of prescription can help guide policy proposals for targeting areas with inadequate antiosteoporotic drug use.

Inequality in oral anticoagulation use and clinical outcomes in atrial fibrillation: a Danish nationwide perspective.

To explore the potential of small-area variation analysis as a tool for identifying unwarranted variation in oral anticoagulation (OAC) use and clinical outcomes in patients with atrial fibrillation (AF) and thereby identify locations with opportunity for improvement in AF care. Based on Danish health care registries, we conducted a nationwide historical cohort study including first-time AF patients with a CHA2DS2-VASc (congestive heart failure, hypertension, age ≥ 75, diabetes, thromboembolism, vascular disease, age 65-74, and sex category) score ≥ 2 between 2007 and 2014 (n = 94482). For each administrative region and municipality, we assessed OAC initiation and persistence as well as the risk of ischaemic stroke, haemorrhagic stroke, and other major bleeding, respectively. In addition, potential temporal changes were examined for all outcomes. Initiation of OAC varied among regions from 49.5% to 62.4%. In patients initiating OAC, the proportion of patients still receiving OAC after 1 year varied from 73.9% to 79.3%. Oral anticoagulation use increased in all regions during the study period, particularly after 2010, but regional variation in OAC initiation persisted. Generally, the regions and municipalities with the highest initiation of OAC also had the highest OAC persistence. The risk of ischaemic stroke and other major bleeding was lower in these regions and municipalities. We found no significant difference between regions in risk of haemorrhagic stroke. Substantial geographical variation in OAC use and clinical outcomes occurs in Denmark demonstrating the potential of small-area variation analysis as a tool for identifying unwarranted variation in AF care and clinical outcomes. Our findings demonstrate the need for additional initiatives to ensure uniform high-quality care for AF patients.

Asymptotic random extremal ratio and product based on generalized order statistics and its dual

ABSTRACTThe extremal ratio has been used in several fields, most notably in industrial quality control, life testing, small-area variation analysis, and the classical heterogeneity of variance situation. In many biological, agricultural, military activity problems and in some quality control problems, it is almost impossible to have a fixed sample size, because some observations are always lost for various reasons. Therefore, the sample size itself is considered frequently to be an random variable (rv). Generalized order statistics (GOS) have been introduced as a unifying theme for several models of ascendingly ordered rvs. The concept of dual generalized order statistics (DGOS) is introduced to enable a common approach to descendingly ordered rvs. In this article, the limit dfs are obtained for the extremal ratio and product with random indices under non random normalization based on GOS and DGOS. Moreover, this article considers the conditions under which the cases of random and non random indices give the same asymptotic results. Some illustrative examples are obtained, which lend further support to our theoretical results.

Prevalence and Incidence of Diagnosed Chronic Rhinosinusitis in Alberta, Canada

Reported prevalence rates of chronic rhinosinusitis (CRS) range from 1% to 12% worldwide. To facilitate appropriate health service delivery and resource allocation, it is important to improve the estimated burden of CRS to the health care system. To assess the prevalence and incidence of diagnosed CRS in Alberta, Canada, from the perspective of the health care system and to evaluate the 10-year temporal trend and geographic variation of diagnosed CRS. From provincial-wide physicians' claim data, a CRS cohort was identified using a validated case definition. The population at the midpoint (2008-2009) of the study period (2 925 930) was used as the reference. The crude as well as age- and sex-standardized incidence and prevalence rates were calculated. The age-specific incidence and prevalence by sex were also assessed in each study year. Small-area variation analysis was conducted using extremal quotient, weighted coefficient of variation, χ2 statistic, systematic component of variation, and empirical Bayes variance estimate. Of the 2 925 930 individuals in the study at midpoint (2008-2009), 1 451 261 (49.6%) were women, and the mean (SD) age was 45 (17) years. From fiscal year 2004-2005 to fiscal year 2013-2014, the mean age- and sex-standardized incidence of diagnosed CRS was 2.5 (range, 2.3-2.7) per 1000 population. The estimated prevalence based on age-specific incidence varied between 18.8 (95% CI, 18.7-18.9) and 23.3 (95% CI, 23.1-23.5) per 1000 population during 2004-2005 to 2013-2014, and no obvious growing trend was found. There was high geographic variation in the diagnosed incidence and prevalence of CRS (mean systematic component of variation, 19.4 and 12.3, respectively). Although the incidence and prevalence rates of diagnosed CRS were lower compared with earlier published estimates obtained from population-based survey analysis, outcomes from this study may more accurately reflect the disease burden of CRS to the health care system. Given that the prevalence of CRS within a single province is expected to be uniformly distributed, the large geographic variation in diagnosed CRS indicates a potential gap in quality of care and justifies further investigation into the reasons for the variation.

Geographic Variation of Endoscopic Sinus Surgery in Canada: An Alberta-Based Small Area Variation Analysis.

With an estimated 10,000 to 15,000 endoscopic sinus surgery (ESS) cases performed in Canada each year, identifying potential unwarranted practice patterns is important. The objective of this study is to examine the rates and geographic variation of ESS in the province of Alberta, Canada. Small area variation analysis. Province of Alberta, Canada. The National Ambulatory Care Reporting System database was searched to identify all patients who received ESS between April 1, 2010, and March 31, 2013, in Alberta, Canada. The annual adjusted rates of ESS per 1000 people were calculated for each Alberta health zone and health status area. Geographic variations were evaluated with the extremal quotient, weighted coefficient of variation, and systematic component of variance. Chi-squared-test was used to quantify the significance of variation of the adjusted ESS rates across regions. The annual adjusted rate of ESS was 0.33 per 1000 people in Alberta, Canada. The mean extremal quotient for health status areas was 6.9, indicating a 7-fold difference between the highest and lowest regions. The mean coefficient of variation was 41.0, and the mean systematic component of variance was 10.5, which demonstrates "very high" variation. This study observed very high geographic variation in the rates of ESS across the province of Alberta. Given the negative impact of unwarranted surgical variation on quality of care, outcomes from this study indicate a need to further evaluate the delivery of care for ESS in Canada to improve overall health system performance.

A review of geographic variation and Geographic Information Systems (GIS) applications in prescription drug use research

While understanding geography's role in healthcare has been an area of research for over 40 years, the application of geography-based analyses to prescription medication use is limited. The body of literature was reviewed to assess the current state of such studies to demonstrate the scale and scope of projects in order to highlight potential research opportunities. To review systematically how researchers have applied geography-based analyses to medication use data. Empiric, English language research articles were identified through PubMed and bibliographies. Original research articles were independently reviewed as to the medications or classes studied, data sources, measures of medication exposure, geographic units of analysis, geospatial measures, and statistical approaches. From 145 publications matching key search terms, forty publications met the inclusion criteria. Cardiovascular and psychotropic classes accounted for the largest proportion of studies. Prescription drug claims were the primary source, and medication exposure was frequently captured as period prevalence. Medication exposure was documented across a variety of geopolitical units such as countries, provinces, regions, states, and postal codes. Most results were descriptive and formal statistical modeling capitalizing on geospatial techniques was rare. Despite the extensive research on small area variation analysis in healthcare, there are a limited number of studies that have examined geographic variation in medication use. Clearly, there is opportunity to collaborate with geographers and GIS professionals to harness the power of GIS technologies and to strengthen future medication studies by applying more robust geospatial statistical methods.

Whoʼs in the Driverʼs Seat? The Influence of Patient and Physician Enthusiasm on Regional Variation in Degenerative Lumbar Spinal Surgery

Cross-sectional population-based study using administrative databases, census data, and surveys of orthopedic/neurosurgeons, family physicians (FPs) and patients in Ontario, Canada. To determine the influence of the enthusiasm of patients, FPs, and surgeons for surgery on the regional variation in surgical rates for degenerative diseases of the lumbar spine (DDLS), such as spinal stenosis and degenerative spondylolisthesis. Rates of surgery and healthcare costs for treating DDLS have been increasing. Regional variation in spinal surgical rates has been observed and it is thought that the enthusiasm of patients and physicians for surgery contributes to this variation. Using population-based administrative databases, we included all patients aged 50 years and older who underwent DDLS surgery (i.e., decompression/laminectomy, fusion) from 2002 to 2006 and calculated standardized utilization rates across counties. We measured regional "enthusiasm for surgery" for surgeons, FPs, and patients, using responses from a province-wide survey. Small-area variation analysis and multivariate Poisson regression models were performed calculating incidence rate ratios (IRRs) controlling for county demographics, socioeconomic measures, prevalence of disease, and community resources. We identified 10,318 DDLS surgeries (mean age 65 years, 50.6% female). Significant regional variation was observed (extremal quotient 5.0, coefficient of variation 28.0). Counties with higher rates of surgery had higher surgeon enthusiasm for surgery (IRR: 1.26, P < 0.013), older (IRR: 2.17, P < 0.0001) male patients (IRR: 1.19, P < 0.0001), lower income (IRR: 0.89, P < 0.0015), more knowledge of official languages (IRR: 1.12, P < 0.0003), and the presence of magnetic resonance imaging scanners (IRR: 1.30, P < 0.004). FP and patient enthusiasm for surgery, physician supply, and prevalence of disease were not statistically associated with higher surgical rates. Prior studies have not addressed the role of patient enthusiasm for surgery. Although patients and FPs had variable enthusiasm for surgery, surgeon enthusiasm was the dominant potentially modifiable factor influencing surgical rates. Prevalence of disease and community resources were not related to surgical rates. Strategies targeting surgeon practices may reduce regional variation in care and improve access disparities.

Variabilidad en la utilización de antihipertensivos entre las zonas básicas de salud de la Comunidad Valenciana

Objectives To estimate consumption of five subgroups of antihypertensive drugs by primary care areas and to analyze its variation. Methods We performed an ecological, descriptive study of antihypertensive consumption in 239 primary care areas in the autonomous region of Valencia in 2005 followed by analysis of the variability observed. The 239 primary care areas were studied by descriptive analysis of dispensation [defined daily dose (DDD) per 1,000 inhabitants/day in pensioners (DDD/1000p/day) and in the active population (DDD/1000a/day)] and standardized consumption ratios. Small-area variation analysis was used to analyze the observed variability. Associations among dispensations of the distinct therapeutic subgroups were also analyzed. Results Overall antihypertensive use in the autonomous region of Valencia in 2005 was 235.6 DDD/1000/day. This consumption was concentrated in pensioners (800 DDD/1000p/day vs. 73 DDD/1000a/day). Consumption of antihypertensive subgroups oscillated from 442 DDD/1000p/day for drugs with action on the renin-angiotensin system to 32 DDD/1000p/day for doxazosin. The active population showed similar patterns. Variation in consumption was moderate, with coefficients of variation from 0.20 to 0.40 (slightly greater for the active population). Associations among dispensations of the different therapeutic subgroups were strong. Conclusions This study shows major variations in the overall consumption of antihypertensive drugs among primary care areas of the autonomous region of Valencia. These results suggest that variation may be associated with problems of underutilization in areas with lower consumption.

Is there much variation in variation? Revisiting statistics of small area variation in health services research

BackgroundThe importance of Small Area Variation Analysis for policy-making contrasts with the scarcity of work on the validity of the statistics used in these studies. Our study aims at 1) determining whether variation in utilization rates between health areas is higher than would be expected by chance, 2) estimating the statistical power of the variation statistics; and 3) evaluating the ability of different statistics to compare the variability among different procedures regardless of their rates.MethodsParametric bootstrap techniques were used to derive the empirical distribution for each statistic under the hypothesis of homogeneity across areas. Non-parametric procedures were used to analyze the empirical distribution for the observed statistics and compare the results in six situations (low/medium/high utilization rates and low/high variability). A small scale simulation study was conducted to assess the capacity of each statistic to discriminate between different scenarios with different degrees of variation.ResultsBootstrap techniques proved to be good at quantifying the difference between the null hypothesis and the variation observed in each situation, and to construct reliable tests and confidence intervals for each of the variation statistics analyzed. Although the good performance of Systematic Component of Variation (SCV), Empirical Bayes (EB) statistic shows better behaviour under the null hypothesis, it is able to detect variability if present, it is not influenced by the procedure rate and it is best able to discriminate between different degrees of heterogeneity.ConclusionThe EB statistics seems to be a good alternative to more conventional statistics used in small-area variation analysis in health service research because of its robustness.

Recent trends in hospitalization for diabetic ketoacidosis in ontario children.

To investigate trends and geographic variation in diabetic ketoacidosis (DKA) hospitalization rates among children in Ontario from 1991 to 1999. Canadian Institute for Health Information (CIHI) data were used to identify 15,872 diabetes-related hospital admissions in children younger than 19 years in Ontario from 1991 to 1999. Of these, 5,008 admissions were because of DKA and 10,864 admissions were because of conditions other than DKA (non-DKA). Small area variation analysis was used to compare areas with high versus low DKA admission rates. There was a 19% relative decrease in the overall diabetes admission rate over the study period. Non-DKA admissions decreased by 29%, whereas DKA admissions remained stable. Total days of care decreased by 393 days per year for non-DKA admissions and by 99 days per year for DKA admissions. The average length of hospital stay decreased from 4.9 to 3.5 days for non-DKA admissions and from 4.5 to 3.2 days for DKA admissions. The fatality rate was 0.19% for non-DKA admissions and 0.18% for DKA admissions. Variation across geographic areas remained stable for DKA over the study period (Kendall's correlation coefficient 0.64, P = 0.017) with an average 3.7-fold difference between the lowest and highest regions. Increased ambulatory care efforts for children with type 1 diabetes in Ontario have successfully reduced non-DKA admission rates. However, DKA admission rates have remained stable. Geographic variation for DKA admissions is low, but the observed 3.7-fold difference is clinically important for a preventable complication with a significant potential for long-term morbidity and mortality. Prevention strategies are needed, particularly in areas identified with the highest rates.

Relationship between rates and outcomes of operative treatment for lumbar disc herniation and spinal stenosis.

Population-based variations in rates of operations for the treatment of lumbar disc herniation and spinal stenosis are well known. This variability may occur in part because of differences in the threshold at which physicians recommend an operation, reflecting uncertainty about the optimum use of an operative procedure. To the best of our knowledge, no previous reports have indicated whether differences in population-based rates of operative treatment are associated with patient outcomes. The Maine Lumbar Spine Study is an ongoing prospective study of 655 patients who had a herniated lumbar disc or spinal stenosis. The patients were enrolled by their physicians, who provided baseline demographic and treatment-related data. The patients completed baseline and follow-up questionnaires that focused on symptoms, function, satisfaction, and quality of life. Small-area variation analysis was used to develop three distinct so-called spine service areas in Maine. The outcomes (usually at four years; minimum, two years) were compared among these areas, in which a total of 250 patients had been managed operatively and had answered questionnaires. Population-based rates of operative treatment derived from statewide data that had been collected over five years in the state of Maine ranged from 38 percent below to 72 percent above the average rate in the state (a greater than fourfold difference). The outcomes for the patients who had been managed by surgeons in the lowest-rate area were superior to those for the patients in the two higher-rate areas. Seventy-nine percent (fifty-seven) of seventy-two patients in the lowest-rate area had marked or complete relief of pain in the lower extremity compared with 60 percent (eighteen) of thirty patients in the highest-rate area. The improvements in the Roland disability score (p < or = 0.01), quality of life (p < or = 0.01), and satisfaction (p < or = 0.05) were significantly greater among the patients in the lowest-rate area. The patients in the higher-rate areas generally had less severe symptoms and findings at baseline than those in the lowest-rate area did. Higher population-based rates of elective spinal operations may be associated with inferior outcomes. This variability is possibly related to differences in physicians' preferences with regard to recommending an operation and in their criteria for the selection of patients. Physicians cannot assume that their outcomes will be the same as those of others, and therefore they need to evaluate their own results.

Small Area Variation Analysis

In small-area variation analysis, the variation of health care utilization rates, e.g., admission rates, among small areas is calculated. Frequently, the variation of one diagnosis, diagnosis-related group (DRG), or procedure is compared with the variation of another. Unfortunately, the methods generally used to make these comparisons are not consistent. They differ on whether they 1) adjust for the prevalence of the DRGs, 2) distinguish between variation among areas and variation within areas, 3) weight all areas equally, and 4) adjust for multiple admissions per person. None has an associated confidence interval. These discrepancies occur in part because there is no statistical model of small area variation. Without such a model, it is not known how to measure variation, and thus, it is not known how to compare different DRGs. Here, the authors use data on 473 DRGs from 28 counties in Washington state to study the nature of variability. The variation was higher for the more prevalent DRGs, suggesting that adjusting for prevalence may be reasonable. The true coefficient of variation appears to be a "natural" measure of variation, but the usual small area variation statistics do not provide good estimates of the true coefficient of variation. A new estimate is proposed that can be used to compare and test the variability of several DRGs.