To study whether neonatal and infant mortality, after adjustments for differences in case mix, were independent of the type of hospital in which the delivery was carried out.
The Medical Birth Registry of Norway provided detailed medical information for all births in Norway.
Hospitals were classified into two groups: local hospitals/maternity clinics versus central/regional hospitals. Outcomes were neonatal and infant mortality. The data were analyzed using propensity score weighting to make adjustments for differences in case mix between the two groups of hospitals. This analysis was supplemented with analyses of 13 local hospitals that were closed. Using a difference-in-difference approach, the effects that these closures had on neonatal and infant mortality were estimated.
Neonatal and infant mortality were not affected by the type of hospital where the delivery took place.
A regionalized maternity service does not lead to increased neonatal and infant mortality. This is mainly because high-risk deliveries were identified well in advance of the birth, and referred to a larger hospital with sufficient perinatal resources to deal with these deliveries.
Program evaluation; risk adjustment for clinical outcomes; determinants of health; obstetrics/gynecology
We examined the concurrent and lagged effects of registered nurse (RN) turnover on unit-acquired pressure ulcer rates and whether RN staffing mediated the effects.
Quarterly unit-level data were obtained from the National Database of Nursing Quality Indicators for 2008 to 2010. A total of 10,935 unit-quarter observations (2,294 units, 465 hospitals) were analyzed.
This longitudinal study used multilevel regressions and tested time-lagged effects of study variables on outcomes.
The lagged effect of RN turnover on unit-acquired pressure ulcers was significant, while there was no concurrent effect. For every 10 percentage-point increase in RN turnover in a quarter, the odds of a patient having a pressure ulcer increased by 4 percent in the next quarter. Higher RN turnover in a quarter was associated with lower RN staffing in the current and subsequent quarters. Higher RN staffing was associated with lower pressure ulcer rates, but it did not mediate the relationship between turnover and pressure ulcers.
We suggest that RN turnover is an important factor that affects pressure ulcer rates and RN staffing needed for high-quality patient care. Given the high RN turnover rates, hospital and nursing administrators should prepare for its negative effect on patient outcomes.
Registered nurse turnover; nurse staffing; inpatient outcomes; pressure ulcers
To determine whether quality measures based on computer-extracted EHR data can reproduce findings based on data manually extracted by reviewers.
We studied 12 measures of care indicated for adolescent well-care visits for 597 patients in three pediatric health systems.
Data Collection/Extraction Methods
Manual reviewers collected quality data from the EHR. Site personnel programmed their EHR systems to extract the same data from structured fields in the EHR according to national health IT standards.
Overall performance measured via computer-extracted data was 21.9 percent, compared with 53.2 percent for manual data. Agreement measures were high for immunizations. Otherwise, agreement between computer extraction and manual review was modest (Kappa = 0.36) because computer-extracted data frequently missed care events (sensitivity = 39.5 percent). Measure validity varied by health care domain and setting. A limitation of our findings is that we studied only three domains and three sites.
The accuracy of computer-extracted EHR quality reporting depends on the use of structured data fields, with the highest agreement found for measures and in the setting that had the greatest concentration of structured fields. We need to improve documentation of care, data extraction, and adaptation of EHR systems to practice workflow.
Quality measurement; pediatric well-care; electronic health records
To estimate the impact of Express Lane Eligible (ELE) implementation on Medicaid/CHIP enrollment in eight states.
Data Sources/Study Setting
2007 to 2011 data from the Statistical Enrollment Data System (SEDS) on Medicaid/CHIP enrollment.
We estimate difference-in-difference equations, with quarter and state fixed effects. The key independent variable is an indicator for whether the state had ELE in place in the given quarter, allowing the experience of statistically matched non-ELE states to serve as a formal counterfactual against which to assess the changes in the eight ELE states. The model also controls for time-varying economic and policy factors within each state.
Data Collection/Extraction Methods
We obtained SEDS enrollment data from CMS.
Across model specifications, the ELE effects on Medicaid enrollment among children were consistently positive, ranging between 4.0 and 7.3 percent, with most estimates statistically significant at the 5 percent level. We also find that ELE increased combined Medicaid/CHIP enrollment.
Our results imply that ELE has been an effective way for states to increase enrollment and retention among children eligible for Medicaid/CHIP. These results also imply that ELE-like policies could improve take-up of subsidized coverage under the ACA.
Evaluation design and research; health economics; program evaluation; state health policies; Express Lane Eligiblity, Medicaid
To better understand the issue of inappropriate pediatric Emergency Department (ED) visits in Italy, including the impact of the last National Health System reform.
A retrospective cohort study was conducted with five health care providers in the Veneto region (Italy) in a 2-year period (2010–2011). ED visits were considered “inappropriate” by evaluating both nursing triage and resource utilization, as addressed by the Italian Ministry of Health in 2007. Factors associated with inappropriate ED visits were identified. The cost of each visit was calculated.
In total, 134,358 ED visits with 455,650 performed procedures were recorded in the 2-year period; of these, 76,680 (57.1 percent) were considered inappropriate ED visits. Patients likely to make inappropriate ED visits were younger, female, visiting the ED during night or holiday, when the primary care provider (PCP) is not available.
The National Health System reform aims to improve efficiency, effectiveness, and costs by opening PCP offices 24 hours a day and 7 days a week. This study highlights the need for a deep reorganization of the Italian Primary Care System not only providing a larger time availability but also treating the parents' lack of education on children's health.
Nonurgent; inappropriate; pediatric emergency department; health expenditure; reorganization
To evaluate the effect of medical home implementation on primary care delivery in the Veterans Health Administration (VHA).
Data Sources/Study Setting/Study Design
We link interview-based qualitative data on medical home implementation to quantitative outcomes from VHA clinical encounter data. We use a longitudinal analysis with provider fixed effects (taking advantage of variation in timing of implementation and allowing each provider to serve as a control for him or herself) to test whether patient-aligned care team (PACT) implementation was associated with changes in organizational processes and patient outcomes.
Among 683 PCPs, caring for 321,295 patients, the uptake of eight of nine PACT structural changes significantly increased from July 2010 to June 2012 as did the percentage of primary care appointments occurring by telephone and hospital discharges contacted within 2 days of discharge. We found that PACT implementation was associated with significant improvements in 2-day post-hospital discharge contact, but not primary care visits occurring by telephone or within 3 days of the requested date. We found no association between medical home implementation and rates of emergency department use by patients.
Medical home implementation at the VHA resulted in large changes in the structure of care but few changes in patient-level outcomes. These results highlight both the complexity of studying the effect of the medical home as well as implementing this model to change primary care delivery.
Medical home; veterans; primary care delivery
To assess whether patient choice of physician or health plan was affected by physician tier-rankings.
Administrative claims and enrollment data on 171,581 nonelderly beneficiaries enrolled in Massachusetts Group Insurance Commission health plans that include a tiered physician network and who had an office visit with a tiered physician.
We estimate the impact of tier-rankings on physician market share within a plan of new patients and on the percent of a physician's patients who switch to other physicians with fixed effects regression models. The effect of tiering on consumer plan choice is estimated using logistic regression and a pre–post study design.
Physicians in the bottom (least-preferred) tier, particularly certain specialist physicians, had lower market share of new patient visits than physicians with higher tier-rankings. Patients whose physician was in the bottom tier were more likely to switch health plans. There was no effect of tier-ranking on patients switching away from physicians whom they have seen previously.
The effect of tiering appears to be among patients who choose new physicians and at the lower end of the distribution of tiered physicians, rather than moving patients to the “best” performers. These findings suggest strong loyalty of patients to physicians more likely to be considered their personal doctor.
Tiered networks; consumer choice; providers; managed care; cost-sharing
To develop and test incident drug user designs for assessing cost savings from statin use in diabetics.
Random 5 percent sample of Medicare beneficiaries, 2006–2008.
Seven-step incident user design to assess impact of statin initiation on subsequent Medicare spending: (1) unadjusted pre/post initiation test; (2) unadjusted difference-in-difference (DID) with comparison series; (3) adjusted DID; (4) propensity score (PS)-matched DID with static and dynamic baseline covariates; (5) PS-matched DID by drug adherence strata; (6) PS-matched DID for high adherers controlling for healthy adherer bias; and (7) replication for ACE-inhibitor/ARB initiators.
Data Collection/Extraction Methods
Subjects with prevalent diabetes and no statin use (January–June 2006) and statin initiation (July 2006–January 2008) compared to nonusers with a random “potential-initiation” month. Monthly Medicare spending tracked 24 months pre- and post-initiation.
Statistically significant savings in Medicare spending were observed beginning 7 months post-initiation for statins and 13 months post-initiation for ACEIs/ARBs. However, these savings were only observed for adherent patients in steps 5 and 6.
Drug initiator designs are more robust to confounding than prevalent user designs in assessing cost-offsets from drug use but still require other adjustments and sensitivity analysis to ensure proper inference.
Drug initiator design; statins; cost-offsets
To examine how different response scales, methods of survey administration, and survey format affect responses to the CAHPS (Consumer Assessment of Healthcare Providers and Systems) Clinician and Group (CG-CAHPS) survey.
A total of 6,500 patients from a university health center were randomly assigned to receive the following: standard 12-page mail surveys using 4-category or 6-category response scales (on CG-CAHPS composite items), telephone surveys using 4-category or 6-category response scales, or four-page mail surveys.
A total of 3,538 patients completed surveys. Composite score means and provider-level reliabilities did not differ between respondents receiving 4-category or 6-category response scale surveys or between 12-page and four-page mail surveys. Telephone respondents gave more positive responses than mail respondents.
We recommend using 4-category response scales and the four-page mail CG-CAHPS survey.
CAHPS; survey methods; patient experience surveys
To compare the probability of experiencing a potentially preventable hospitalization (PPH) between older dual eligible Medicaid home and community-based service (HCBS) users and nursing home residents.
Three years of Medicaid and Medicare claims data (2003–2005) from seven states, linked to area characteristics from the Area Resource File.
A primary diagnosis of an ambulatory care sensitive condition on the inpatient hospital claim was used to identify PPHs. We used inverse probability of treatment weighting to mitigate the potential selection of HCBS versus nursing home use.
The most frequent conditions accounting for PPHs were the same among the HCBS users and nursing home residents and included congestive heart failure, pneumonia, chronic obstructive pulmonary disease, urinary tract infection, and dehydration. Compared to nursing home residents, elderly HCBS users had an increased probability of experiencing both a PPH and a non-PPH.
HCBS users’ increased probability for potentially and non-PPHs suggests a need for more proactive integration of medical and long-term care.
Long-term care; home care/nursing homes; acute inpatient care; Medicaid; Medicare
To explore the impact of the Hospital Readmissions Reduction Program (HRRP) on hospitals serving vulnerable populations.
Data Sources/Study Setting
Medicare inpatient claims to calculate condition-specific readmission rates. Medicare cost reports and other sources to determine a hospital's share of duals, profit margin, and characteristics.
Regression analyses and projections were used to estimate risk-adjusted readmission rates and financial penalties under the HRRP. Findings were compared across groups of hospitals, determined based on their share of duals, to assess differential impacts of the HRRP.
Both patient dual-eligible status and a hospital's dual-eligible share of Medicare discharges have a positive impact on risk-adjusted hospital readmission rates. Under current Centers for Medicare and Medicaid Service methodology, which does not adjust for socioeconomic status, high-dual hospitals are more likely to have excess readmissions than low-dual hospitals. As a result, HRRP penalties will disproportionately fall on high-dual hospitals, which are more likely to have negative all-payer margins, raising concerns of unintended consequences of the program for vulnerable populations.
Policies to reduce hospital readmissions must balance the need to ensure continued access to quality care for vulnerable populations.
Readmissions; Medicare; Medicaid; dual eligibles; Affordable Care Act
To explore the relative efficiency of dialysis facilities in the United States and identify factors that are associated with efficiency in the production of dialysis treatments.
Data Sources/Study Setting
Medicare cost report data from 4,343 free-standing dialysis facilities in the United States that offered in-center hemodialysis in 2010.
A cross-sectional, facility-level retrospective database analysis, utilizing data envelopment analysis (DEA) to estimate facility efficiency.
Data Collection/Extraction Methods
Treatment data and cost and labor inputs of dialysis treatments were obtained from 2010 Medicare Renal Cost Reports. Demographic data were obtained from the 2010 U.S. Census.
Only 26.6 percent of facilities were technically efficient. Neither the intensity of market competition nor the profit status of the facility had a significant effect on efficiency. Facilities that were members of large chains were less likely to be efficient. Cost and labor savings due to changes in drug protocols had little effect on overall dialysis center efficiency.
The majority of free-standing dialysis facilities in the United States were functioning in a technically inefficient manner. As payment systems increasingly employ capitation and bundling provisions, these institutions will need to evaluate their efficiency to remain competitive.
Dialysis market; efficiency; data envelopment analysis
To determine the impact of Florida's Medicaid Demonstration 4 years post-implementation on per member per month (PMPM) Medicaid expenditures and whether receiving care through HMOs versus provider service networks (PSNs) in the Demonstration was associated with PMPM expenditures.
Florida Medicaid claims from two fiscal years prior to implementation of the Demonstration (FY0405, FY0506) and the first four fiscal years after implementation (FY0607-FY0910) from two urban Demonstration counties and two urban non-Demonstration counties.
A difference-in-difference approach was used to compare changes in enrollee expenditures before and after implementation of the Demonstration overall and specifically for HMOs and PSNs.
Claims data were extracted for enrollees in the Demonstration and non-Demonstration counties and collapsed into monthly amounts (N = 26,819,987 person-months).
Among SSI enrollees, the Demonstration resulted in lower increases in PMPM expenditures over time ($40) compared with the non-Demonstration counties ($186), with Demonstration PSNs lowering PMPM expenditures by $7 more than HMOs. Savings were also seen among TANF enrollees but to a lesser extent.
The Medicaid Demonstration in Florida appears to result in lower PMPM expenditures. Demonstration PSNs generated slightly greater reductions in expenditures compared to Demonstration HMOs. PSNs appear to be a promising model for delivering care to Medicaid enrollees.
Administrative data uses; health care costs; Medicaid; health care organizations and systems
To examine the relationship between insurance market structure and health care prices, utilization, and spending.
Claims for 37.6 million privately insured employees and their dependents from the Truven Health Market Scan Database in 2009. Measures of insurer market structure derived from Health Leaders Inter study data.
Regression models are used to estimate the association between insurance market concentration and health care spending, utilization, and price, adjusting for differences in patient characteristics and other market-level traits.
Insurance market concentration is inversely related to prices and spending, but positively related to utilization. Our results imply that, after adjusting for input price differences, a market with two equal size insurers is associated with 3.9 percent lower medical care spending per capita (p = .002) and 5.0 percent lower prices for health care services relative to one with three equal size insurers (p < .001).
Greater fragmentation in the insurance market might lead to higher prices and higher spending for care, suggesting some of the gains from insurer competition may be absorbed by higher prices for health care. Greater attention to prices and utilization in the provider market may need to accompany procompetitive insurance market strategies.
Geographic variation; spending; utilization; competition; markets
Patients are treated using observation services (OS) when their care needs exceed standard outpatient care (i.e., clinic or emergency department) but do not qualify for admission. Medicare and other private payers seek to limit this care setting to 48 hours.
Data Source/Study Setting
Healthcare Cost and Utilization Project data from 10 states and data collected from two additional states for 2009.
Bivariate analyses and hierarchical linear modeling were used to examine patient- and hospital-level predictors of OS stays exceeding 48 (and 72) hours (prolonged OS). Hierarchical models were used to examine the additional cost associated with longer OS stays.
Of the 696,732 patient OS stays, 8.8 percent were for visits exceeding 48 hours. Having Medicaid or no insurance, a condition associated with no OS treatment protocol, and being discharged to skilled nursing were associated with having a prolonged OS stay. Among Medicare patients, the mean charge for OS stays was $10,373. OS visits of 48–72 hours were associated with a 42 percent increase in costs; visits exceeding 72 hours were associated with a 61 percent increase in costs.
Patient cost sharing for most OS stays of less than 24 hours is lower than the Medicare inpatient deductible. However, prolonged OS stays potentially increase this cost sharing.
Observation services; observation unit; patient cost; Medicare payment policy
To examine how enrollees' statin compliance responds to expected prices in Medicare Part D, which features a nonlinear price schedule due to a coverage gap.
Data Sources/Study Setting
Prescription Drug Event data for a 5 percent random sample of Medicare Advantage Prescription Drug Plan enrollees in 2008 who did not receive a low-income subsidy.
We analyze statin compliance prior to the coverage gap, where the “effective price” is higher than the actual copayment for drugs because consumers anticipate that more spending will make them more likely to reach the gap. We construct each enrollee's effective price as her expected price at the end of the year, which is the weighted average between pre-gap and in-gap copayments with the weight being the predicted probability of hitting the gap. Compliance is defined as at least 80 percent of days covered.
Part D enrollees' pre-gap statin compliance decreases by 3.7–4.7 percentage points for a $10 increase in the effective price.
The presence of a coverage gap decreases statin compliance prior to the gap, suggesting that incorporating expected future prices is important to assess the full impact of cost sharing on drug compliance under nonlinear price schedules.
Nonlinear pricing; statin compliance; Medicare part D
To provide new evidence on whether and how patterns of health care utilization deviate from horizontal equity in a country with a universal and egalitarian public health care system: Italy.
Secondary analysis of data from the Health Conditions and Health Care Utilization Survey 2005, conducted by the Italian National Institute of Statistics on a probability sample of the noninstitutionalized Italian population.
Using multilevel logistic regression, we investigated how the probability of utilizing five health care services varies among individuals with equal health status but different SES.
Respondents aged 18 or older at the interview time (n = 103,651).
Overall, we found that use of primary care is inequitable in favor of the less well-off, hospitalization is equitable, and use of outpatient specialist care, basic medical tests, and diagnostic services is inequitable in favor of the well-off. Stratifying the analysis by health status, however, we found that the degree of inequity varies according to health status.
Despite its universal and egalitarian public health care system, Italy exhibits a significant degree of SES-related horizontal inequity in health services utilization.
Health care utilization; horizontal equity; SES inequity; Italy
To identify the impact of the Health Center Growth Initiative on access to care for low-income adults.
Data on federal funding for health centers are from the Bureau of Primary Health Care's Uniform Data System (2000–2007), and individual-level measures of access and use are derived from the National Health Interview Survey (2001–2008).
We estimate person-level models of access and use as a function of individual- and market-level characteristics. By using market-level fixed effects, we identify the effects of health center funding on access using changes within markets over time. We explore effects on low-income adults and further examine how those effects vary by insurance coverage.
We calculate health center funding per poor person in a health care market and attach this information to individual observations on the National Health Interview Survey. Health care markets are defined as hospital referral regions.
Low-income adults in markets with larger funding increases were more likely to have an office visit and to have a general doctor visit. These results were stronger for uninsured and publicly insured adults.
Expansions in federal health center funding had some mitigating effects on the access declines that were generally experienced by low-income adults over this time period.
Health centers; access to care; primary care; safety net
To understand patient loyalty to providers over time, informing effective population health management.
Patient care-seeking patterns over a 6-year timeframe in Minnesota, where care systems have a significant portion of their revenue generated by shared-saving contracts with public and private payers.
Weibull duration and probit models were used to examine patterns of patient attribution to a care system and the continuity of patient affiliation with a care system. Clustering of errors within family unit was used to account for within-family correlation in unobserved characteristics that affect patient loyalty.
The payer provided data from health plan administrative files, matched to U.S. Census-based characteristics of the patient's neighborhood. Patients were retrospectively attributed to health care systems based on patterns of primary care.
I find significant patient loyalty, with past loyalty a very strong predictor of future relationship. Relationships were shorter when the patient's health status was complex and when the patient's care system was smaller.
Population health management can be beneficial to the care system making this investment, particularly for patients exhibiting prior continuity in care system choice. The results suggest that co-located primary and specialty services are important in maintaining primary care loyalty.
Consumer issues; physician payment; organization and delivery of care
This study examined methodological concerns with standard approaches to measuring race and ethnicity using the federally defined race and ethnicity categories, as utilized in National Institutes of Health (NIH) funded research.
Data Sources/Study Setting
Surveys were administered to 219 economically disadvantaged, racially and ethnically diverse participants at Boston Women Infants and Children (WIC) clinics during 2010.
We examined missingness and misclassification in responses to the closed-ended NIH measure of race and ethnicity compared with open-ended measures of self-identified race and ethnicity.
Rates of missingness were 26 and 43 percent for NIH race and ethnicity items, respectively, compared with 11 and 18 percent for open-ended responses. NIH race responses matched racial self-identification in only 44 percent of cases. Missingness and misclassification were disproportionately higher for self-identified Latina(o)s, African-Americans, and Cape Verdeans. Race, but not ethnicity, was more often missing for immigrant versus mainland U.S.-born respondents. Results also indicated that ethnicity for Hispanic/Latina(o)s is more complex than captured in this measure.
The NIH's current race and ethnicity measure demonstrated poor differentiation of race and ethnicity, restricted response options, and lack of an inclusive ethnicity question. Separating race and ethnicity and providing respondents with adequate flexibility to identify themselves both racially and ethnically may improve valid operationalization.
Measurement of race and ethnicity; health disparities research; National Institutes of Health (NIH) race and ethnicity reporting; racial and ethnic self-identification
Examine how widely used statistical benchmarks of health care provider performance compare with histogram-based statistical benchmarks obtained via hierarchical Bayesian modeling.
Publicly available data from 3,240 hospitals during April 2009–March 2010 on two process-of-care measures reported on the Medicare Hospital Compare website.
Secondary data analyses of two process-of-care measures comparing statistical benchmark estimates and threshold exceedance determinations under various combinations of hospital performance measure estimates and benchmarking approaches.
Statistical benchmarking approaches for determining top 10 percent performance varied with respect to which hospitals exceeded the performance benchmark; such differences were not found at the 50 percent threshold. Benchmarks derived from the histogram of provider performance under hierarchical Bayesian modeling provide a compromise between benchmarks based on direct (raw) estimates, which are overdispersed relative to the true distribution of provider performance and prone to high variance for small providers, and posterior mean provider performance, for which over-shrinkage and under-dispersion relative to the true provider performance distribution is a concern.
Given the rewards and penalties associated with characterizing top performance, the ability of statistical benchmarks to summarize key features of the provider performance distribution should be examined.
Bayesian statistics; hierarchical model; provider profiling; public reporting; statistical benchmark
In their recent Health Services Research article titled “Squeezing the Balloon: Propensity Scores and Unmeasured Covariate Balance,” Brooks and Ohsfeldt (2013) addressed an important topic on the balancing property of the propensity score (PS) with respect to unmeasured covariates. They concluded that PS methods that balance measured covariates between treated and untreated subjects exacerbate imbalance in unmeasured covariates that are unrelated to measured covariates. Furthermore, they emphasized that for PS algorithms, an imbalance on unmeasured covariates between treatment and untreated subjects is a necessary condition to achieve balance on measured covariates between the groups. We argue that these conclusions are the results of their assumptions on the mechanism of treatment allocation. In addition, we discuss the underlying assumptions of PS methods, their advantages compared with multivariate regression methods, as well as the interpretation of the effect estimates from PS methods.