[Show abstract][Hide abstract] ABSTRACT: A host of recent research has used the reweighting methods developed by DiNardo, Fortin, and Lemieux (1996) to analyze the extent to which between-group differences in outcomes are predicted by differences in the distribution of observable characteristics. Much less attention has been paid to isolating the roles of individual covariates in reweighting estimation. We analyze two approaches that have been used in previous studies, and we propose a new approach that can be viewed as a generalization of regression-based methods. We illustrate the differences between the methods using two Monte Carlo designs and an empirical analysis of black-white gaps in birth weight and infant mortality.
Journal of Applied Econometrics 01/2015; · 1.76 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Objectives. We examined how changes in demographic, geographic, and childbearing risk factors were related to changes in the Black-White infant mortality rate (IMR) gap over 2 decades. Methods. Using 1983-2004 Vital Statistics, we applied inverse probability weighting methods to examine the relationship between risk factors and 3 outcomes: the overall IMR gap, its birth weight component, and its conditional (on birth weight) IMR component. Results. The unexplained IMR gap (the part not related to observed risk factors) was stable, changing from 5.0 to 5.3 deaths per 1000 live births. By contrast, the explained gap declined from 4.6 to 1.9. The decline in the explained gap was driven by the changing relationship between risk factors and IMR. Further analysis revealed that most of the unexplained gap occurred among infants weighing less than 1000 grams at birth, whereas most of the explained gap occurred among infants weighing more than 1000 grams. Conclusions. The unexplained gap was stable over the last 2 decades, but the explained gap declined markedly. If the stability of the unexplained gap continues, even complete convergence of risk factors would reduce the Black-White IMR gap by only one quarter. (Am J Public Health. Published online ahead of print December 19, 2013: e1-e7. doi:10.2105/AJPH.2013.301349).
American Journal of Public Health 12/2013; · 3.93 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: The goal was to compare educational attainment and labor market outcomes in young adulthood (21-26 years of age) for a Canadian, population-based cohort of 149 extremely low birth weight (ELBW) (<1000 g) survivors and a normal birth weight (NBW) cohort of 133 young adults from the same geographic area who were matched to the ELBW cohort in childhood.
We estimated the effects of ELBW status, according to gender, on continuous outcomes through least-squares regression and those on binary outcomes through logistic regression. We controlled for family background and considered neurosensory impairment and IQ as mediating variables.
Controlling for family background, ELBW male subjects were less likely to complete high school or to attend a university than were their NBW counterparts, and their educational attainment was reduced by >1 year. Among subjects who were working, weekly earnings were ∼27% lower. ELBW female effects on education were not significant, but ELBW female subjects were less likely than NBW subjects to be employed or in school and they also seemed to experience lower earnings.
Our findings suggested that ELBW survivors are somewhat less productive as adults, on average, than are subjects born NBW and that effects are not confined to subjects with severe neurosensory impairments. In accord with other studies, however, we found that productivity deficits for most ELBW subjects were not large.
[Show abstract][Hide abstract] ABSTRACT: Infant mortality rates (IMRs) vary widely across racial and ethnic groups in the U.S. for reasons that are not well understood. For example, the large black-white disparity is well known and heavily studied. Over the 2000-03 period, the IMR for non-Hispanic blacks was 13.5 per 1000 live births, while for non-Hispanic whites it was 5.7. Only about one-third of this gap can be accounted for by differences across the groups in characteristics such as mother’s age, education, marital status and early use of prenatal care. Over this same period the IMR was 8.8 among American Indians and Alaskan natives (for brevity, Native Americans), 8.2 among those of Puerto Rican heritage, but only 5.3 among those of Mexican heritage, another economically disadvantaged group. In pairwise comparisons, each of these group IMRs is statistically different from the others, except for the Native American/ Puerto Rican pair.This paper uses linked Birth-Infant Death Vital Statistics data for 2000-03 (with sample sizes ranging from 153,000 births for Native Americans to 9.3 million births for non-Hispanic whites, and including a wealth of background information captured on birth certificates) to explore the factors associated with these disparate disparities. Some striking patterns are already apparent from preliminary work. For example, while the high black IMR can be almost entirely “explained” by the relatively high proportion of births in the left tail of the birth weight and gestational age distributions among blacks, birth weight differences account for only 7 percent of the IMR gap between Native Americans and whites. The gap in neonatal mortality rate (the first 28 days of life) between blacks and whites is 10 times as large as between Native Americans and whites, while for post-neonatal mortality (the remainder of the first year) the Native American/white gap is slightly larger than the black/white gap. Puerto Rican-Americans have a similar IMR to Native Americans, but their birth weight distribution more closely resembles that of blacks.The patterns noted above suggest that very different processes influence the infant mortality of these different groups, all of which are economically disadvantaged compared to whites. Our prior work on the black-white IMR gap indicates that post-neonatal death rates are more closely linked to background characteristics than are neonatal deaths, an investigation we now extend to other groups. We expect that a closer examination of why blacks and Puerto Rican-Americans produce less viable infants (based on characteristics at birth) than Native Americans and Mexican-Americans will further our understanding of all disparities in IMR.
[Show abstract][Hide abstract] ABSTRACT: We provide a unified assessment of a striking disparity in the United States: the differential rate at which white and black infants die. We separate the overall mortality gap into three temporal componentsâ€”fitness at birth, conditional neonatal mortality, and conditional post-neonatal mortalityâ€”and quantify the extent to which each of the components can be predicted using a flexible reweighting method. Almost 90 percent of the overall mortality gap is due to differential fitness at birth, little of which can be predicted by racial differences in background characteristics. The remaining mortality gap stems from conditional post-neonatal mortality differences, nearly all of which can be predicted by background characteristics. The predictability of the mortality gap has declined substantially over the past two decades, largely because the mortality gap among extremely low-fitness infants is increasingly unrelated to background characteristics.
[Show abstract][Hide abstract] ABSTRACT: To determine whether a Nurse-Community Health Worker (CHW) home visiting team, in the context of a Medicaid enhanced prenatal/postnatal services (EPS), would demonstrate greater reduction of depressive symptoms and stress and improvement of psychosocial resources (mastery, self-esteem, social support) when compared with usual Community Care (CC) that includes Medicaid EPS delivered by professionals. Greatest program benefits were expected for women who reported low psychosocial resources, high stress, or both at the time of enrollment. Medicaid eligible pregnant women (N = 613) were randomly assigned to either usual CC or the Nurse-CHW team. Mixed effects regression was used to analyze up to five prenatal and postnatal psychosocial assessments. Compared to usual CC, assignment to the Nurse-CHW team resulted in significantly fewer depressive symptoms, and as hypothesized, reductions in depressive symptoms were most pronounced for women with low psychosocial resources, high stress, or both high stress and low resources. Outcomes for mastery and stress approached statistical significance, with the women in the Nurse-CHW group reporting less stress and greater mastery. Women in the Nurse-CHW group with low psychosocial resources reported significantly less perceived stress than women in usual CC. No differences between the groups were found for self-esteem and social support. A Nurse-CHW team approach to EPS demonstrated advantage for alleviating depressive symptoms in Medicaid eligible women compared to CC, especially for women at higher risk.
Archives of Women s Mental Health 07/2009; 12(6):379-91. · 2.01 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We analyze four methods to measure unexplained gaps in mean outcomes: three decompositions based on the seminal work of Oaxaca (1973) and Blinder (1973) and an approach involving a seemingly naïve regression that includes a group indicator variable. Our analysis yields two principal findings. We show that the coefficient on a group indicator variable from an OLS regression is an attractive approach for obtaining a single measure of the unexplained gap. We also show that a commonly-used pooling decomposition systematically overstates the contribution of observable characteristics to mean outcome differences when compared to OLS regression, therefore understating unexplained differences. We then provide three empirical examples that explore the practical importance of our analytic results.
[Show abstract][Hide abstract] ABSTRACT: This study sought to determine whether an intervention for patients with medically unexplained symptoms in primary care reduced total costs, components of cost, and longer-term costs and whether it led to decreased service use outside the health maintenance organization (HMO).
A randomized controlled trial involving 206 patients with medically unexplained symptoms was conducted in a staff-model HMO. The protocol emphasized the provider-patient relationship and included cognitive-behavioral therapy and pharmacological management. Cost data for medical treatments were derived from the HMO's electronic database. Patients were interviewed about work days lost and out-of-pocket expenses for medical care outside the HMO.
The difference in total costs ($1,071) for the 12-month intervention was not significant. The treatment group had significantly higher costs for antidepressants than the usual-care group ($192 higher) during the intervention, and a larger proportion received antidepressants. The intervention group used less medical care outside the HMO and missed one less work day per month on average (1.23 days), indicating a slight improvement in productivity, but the difference was not significant. The between-group difference in estimated total cost was smaller in the year after the intervention (difference of $341) but were not significant.
The total costs for the intervention group were not significantly different, but the group had greater use of antidepressants. Coupled with findings of improved mental health outcomes for this group in a previous study, the results indicate that the intervention may be cost-effective. The longer-term impact needs to be further studied.
[Show abstract][Hide abstract] ABSTRACT: The objective of this study was to compare the current health status, physical ability, functional limitations, and health care use of extremely low birth weight and normal birth weight young adults.
A longitudinal study was conducted of a population-based cohort of 166 extremely low birth weight survivors (501-1000 g birth weight; 1977-1982 births) and a group of 145 sociodemographically comparable normal birth weight individuals. Current health status, history of illnesses, hospitalizations, use of health resources, and physical self-efficacy were assessed through questionnaires that were administered to the young adults by masked interviewers.
Individuals completed the assessments at a mean age of 23 years. Neurosensory impairments were identified in 27% of extremely low birth weight and 2% of normal birth weight individuals. No differences were reported in the current health status for physical or mental summary scores. Extremely low birth weight young adults reported a higher prevalence of chronic health conditions in the past 6 months. A significantly higher proportion of extremely low birth weight individuals had functional limitations in seeing, hearing, and dexterity and experienced clumsiness and learning difficulties. Except for prescription glasses, medications for depression, and home-care services for extremely low birth weight individuals, there were no significant differences between groups in use of health care resources. Extremely low birth weight individuals had significantly weaker hand grip strength and lower scores for physical self-efficacy, perceived physical ability, and physical self-confidence.
Extremely low birth weight young adults seem to enjoy similar current health status to their normal birth weight peers. However, they continue to have significantly poorer physical abilities and a higher prevalence of chronic health conditions and functional limitations. Contrary to expectations, they do not pose a significant burden to the health care system at young adulthood.
[Show abstract][Hide abstract] ABSTRACT: The goals were to compare the self-reported, health-related quality of life of former extremely low birth weight and normal birth weight infants at young adulthood and to determine whether there were any changes over time.
A prospective, longitudinal, population-based study with concurrent control subjects was performed. We interviewed 143 of 166 extremely low birth weight survivors (birth weight: 501-1000 g; 1977-1982 births) and 130 of 145 sociodemographically comparable, normal birth weight, reference subjects. Neurosensory impairments were present for 27% extremely low birth weight and 2% normal birth weight young adults. Health Utilities Index 2 was used to assess health status, and standard gamble technique was used to measure directly the self-reported, health-related, quality of life and 4 hypothetical health states.
Extremely low birth weight young adults reported more functional limitations in cognition, sensation, mobility, and self-care, compared with control subjects. There were no differences between groups in the mean self-reported, health-related, quality of life or between impaired (n = 38) and nonimpaired (n = 105) extremely low birth weight subjects. However, with a conservative approach of assigning a score of 0 for 10 severely disabled, extremely low birth weight subjects, the mean health-related quality of life was significantly lower than control values. Repeated-measures analysis of variance to compare health-related quality-of-life measurements obtained for young adults and teens showed the same decline in scores over time for both groups. There were no differences between groups in the ratings provided for the hypothetical health states.
At young adulthood, health-related quality of life was not related to size at birth or to the presence of disability. There was a small decrease in health-related quality-of-life scores over time for both groups.
[Show abstract][Hide abstract] ABSTRACT: There is no proven primary care treatment for patients with medically unexplained symptoms (MUS). We hypothesized that a long-term, multidimensional intervention by primary care providers would improve MUS patients' mental health.
HMO in Lansing, MI.
Patients from 18 to 65 years old with 2 consecutive years of high utilization were identified as having MUS by a reliable chart rating procedure; 206 subjects were randomized and 200 completed the study.
From May 2000 to January 2003, 4 primary care clinicians deployed a 12-month intervention consisting of cognitive-behavioral, pharmacological, and other treatment modalities. A behaviorally defined patient-centered method was used by clinicians to facilitate this treatment and the provider-patient relationship.
The primary endpoint was an improvement from baseline to 12 months of 4 or more points on the Mental Component Summary of the SF-36.
Two hundred patients averaged 13.6 visits for the year preceding study. The average age was 47.7 years and 79.1% were females. Using intent to treat, 48 treatment and 34 control patients improved (odds ratio [OR]=1.92, 95% confidence interval [CI]: 1.08 to 3.40; P=.02). The relative benefit (relative "risk" for improving) was 1.47 (CI: 1.05 to 2.07), and the number needed to treat was 6.4 (95% CI: 0.89 to 11.89). The following baseline measures predicted improvement: severe mental dysfunction (P<.001), severe body pain (P=.039), nonsevere physical dysfunction (P=.003), and at least 16 years of education (P=.022); c-statistic=0.75.
The first multidimensional intervention by primary care clinicians led to clinically significant improvement in MUS patients.
Journal of General Internal Medicine 08/2006; 21(7):671-7. · 3.42 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Traditionally, educational attainment, getting a job, living independently, getting married, and parenthood have been considered as markers of successful transition to adulthood.
To describe and compare the achievement and the age at attainment of the above markers between extremely low-birth-weight (ELBW) and normal birth-weight (NBW) young adults.
A prospective, longitudinal, population-based study in central-west Ontario, Canada, of 166 ELBW participants who weighed 501 to 1000 g at birth (1977-1982) and 145 sociodemographically comparable NBW participants assessed at young adulthood (22-25 years). Interviewers masked to participant status administered validated questionnaires via face-to-face interviews between January 1, 2002, and April 30, 2004.
Markers of successful transition to adulthood, including educational attainment, student and/or worker role, independent living, getting married, and parenthood.
At young adulthood, 149 (90%) of 166 ELBW participants and 133 (92%) of 145 NBW participants completed the assessments at mean (SD) age of 23.3 (1.2) years and 23.6 (1.1) years, respectively. We included participants with neurosensory impairments (ELBW vs NBW: 40 [27%] vs 3 [2%]) and 7 proxy respondents. The proportion who graduated from high school was similar (82% vs 87%, P = .21). Overall, no statistically significant differences were observed in the education achieved to date. A substantial proportion of both groups were still pursuing postsecondary education (47 [32%] vs 44 [33%]). No significant differences were observed in employment/school status; 71 (48%) ELBW vs 76 (57%) NBW young adults were permanently employed (P = .09). In a subanalysis, a higher proportion of ELBW young adults were neither employed nor in school (39 [26%] vs 20 [15%], P = .02 by Holm's correction); these differences did not persist when participants with disabilities were excluded. No significant differences were found in the proportion living independently (63 [42%] vs 70 [53%], P = .19), married/cohabitating (34 [23%] vs 33 [25%], P = .69), or who were parents (16 [11%] vs 19 [14%], P = .36). The age at attainment of the above markers was similar for both cohorts.
Our study results indicate that a significant majority of former ELBW infants have overcome their earlier difficulties to become functional young adults.
JAMA The Journal of the American Medical Association 03/2006; 295(6):667-75. · 29.98 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We show that, if the utility function is non-homothetic, environmental taxes can have positive non-environmental effects. These effects are illustrated with specific reference to taxes on gasoline and tobacco, in the context of a computational model. We also clarify the relationship between the “double dividend” (associated with a marginal change from a tax system with low reliance on environmentally motivated taxes) and the situation in which the optimal environmental tax rate is greater than the Pigouvian tax rate. These two situations are generated by rather similar combinations of parameters. Copyright Springer Science + Business Media, Inc. 2005
International Tax and Public Finance 03/2005; 12(2):115-130. · 0.54 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Investigators and clinicians almost always rely on Diagnostic and Statistical Manual of Mental Disorder, 4th edition's (DSM-IV) somatoform disorders (and its derivative diagnoses) to characterize and identify patients with medically unexplained symptoms (MUS). Our objective was to evaluate this use by determining the prevalence of DSM-IV somatoform and nonsomatoform disorders in patients with MUS proven by a gold standard chart review.
In a community-based staff model HMO, we identified subjects for a clinical trial using a systematic and reliable chart rating procedure among high-utilizing MUS patients. Only baseline data are reported here. The World Health Organization Composite International Diagnostic Interview provided full and abridged DSM-IV diagnoses. Patients with full or abridged DSM-IV somatoform diagnoses were labeled "DSM somatoform-positive," whereas those without them were labeled "DSM somatoform-negative."
Two hundred six MUS patients averaged 13.6 visits in the year preceding study, 79.1% were females, and the average age was 47.7 years. We found that 124 patients (60.2%) had a nonsomatoform ("psychiatric") DSM-IV diagnosis of any type; 36 (17.5%) had 2 full nonsomatoform diagnoses, and 41 (19.9%) had >2; 92 (44.7%) had some full anxiety diagnosis and 94 (45.6%) had either full depression or minor depression diagnoses. However, only 9 of 206 (4.4%) had any full DSM-IV somatoform diagnosis, and only 39 (18.9%) had abridged somatization disorder. Thus, 48 (23.3%) were "DSM somatoform-positive" and 158 (76.7%) were "DSM somatoform-negative." The latter exhibited less anxiety, depression, mental dysfunction, and psychosomatic symptoms (all p <.001) and less physical dysfunction (p = .011). Correlates of this DSM somatoform-negative status were female gender (p = .007), less severe mental (p = .007), and physical dysfunction (p = .004), a decreased proportion of MUS (p <.10), and less psychiatric comorbidity (p <.10); c-statistic = 0.77.
We concluded that depression and anxiety characterized MUS patients better than the somatoform disorders. Our data suggested radically revising the somatoform disorders for DSM-V by incorporating a new, very large group of now-overlooked DSM somatoform-negative patients who were typically women with less severe dysfunction.
Psychosomatic Medicine 01/2005; 67(1):123-9. · 4.09 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: There are no proven, comprehensive treatments in primary care for patients with medically unexplained symptoms (MUS) even though these patients have high levels of psychosocial distress, medical disability, costs, and utilization. Despite extensive care, these common patients often become worse.
We sought to identify an effective, research-based treatment that can be conducted by primary care personnel.
We used our own experiences and files, consulted with experts, and conducted an extensive review of the literature to identify two things: 1). effective treatments from randomized controlled trials for MUS patients in primary care and in specialty settings; and 2). any type of treatment study in a related area that might inform primary care treatment, for example, depression, provider-patient relationship.
We developed a multidimensional treatment plan by integrating several areas of the literature: collaborative/stepped care, cognitive-behavioral treatment, and the provider-patient relationship. The treatment is designed for primary care personnel (physicians, physician assistants, nurse practitioners) and deployed intensively at the outset; visit intervals are progressively increased as stability and improvement occur.
Providing a comprehensive treatment plan for chronic, high-utilizing MUS patients removes one barrier to treating this common problem effectively in primary care by primary care personnel.
Journal of General Internal Medicine 07/2003; 18(6):478-89. · 3.42 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Patients with medically unexplained symptoms (MUS) often are a source of frustration for clinicians, and despite high quality biomedical attention and frequent diagnostic tests, they have poor health outcomes. Following upon progress in depression treatment approaches, we developed a multidimensional treatment protocol for deployment by primary care personnel. This multi-faceted intervention for MUS patients emphasized cognitive-behavioral principles, the provider-patient relationship, pharmacological management, and treating comorbid medical diseases. We deployed it in an HMO using nurse practitioners (NP) to deliver the intervention to 101 patients, while 102 controls continued to receive medical care from their usual primary care physician. Successful deployment of the intervention required training the NPs, continuing support for the NPs in their management of this difficult population, and establishing strong communication links with the HMO. This paper addresses the practical considerations of using primary care personnel to implement a complex intervention in primary care, and it includes a discussion of special challenges encountered as well as solutions developed to overcome them.
General Hospital Psychiatry 03/2003; 25(2):63-73. · 2.90 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Our objectives were to determine how patients who make frequent use of the medical system (high users) with medically unexplained symptoms met our chart-rating criteria for somatization and minor acute illness and what the stability of such diagnoses were over time.
A chart review was performed at baseline and 1 and 2 years; we re-rated the charts of patients initially rated as having somatization, as well as a 15% sample of those with minor acute illness.
We obtained a random sample of high-use patients (= 6 visits/year) aged 21 to 55 years who were identified from the management information system.
We measured chart review designations as organic disease, somatization, or minor acute illness.
Among 883 high users at baseline, 35% had organic diseases; 14% had somatization; and 51% had minor acute illness as their primary problems. No patients with initial minor acute diagnoses were reclassified as having somatization 1 or 2 years later, and all but 2 patients had minor acute illness in 1 or both follow-up years.
Minor acute illness was more common among high users than somatization and organic diseases combined. It has not previously been studied but probably has been recognized by clinicians as the "worried well." Diagnoses of somatization were unstable over 2 years follow-up, while minor acute diagnoses were stable, supporting the latter as a valid entity.
The Journal of family practice 02/2002; 51(1):24-9. · 0.74 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Somatization is a common, costly problem with great morbidity, but there has been no effective screening method to identify these patients and target them for treatment.
We tested a hypothesis that we could identify high utilizing somatizing patients from a management information system (MIS) by total number of visits and what we termed "somatization potential," the percentage of visits for which ICD-9 primary diagnosis codes represented disorders in the musculoskeletal, nervous, or gastrointestinal systems or ill-defined complaints.
We identified 883 high users from the MIS of a large staff model HMO as those having six or more visits during the year studied (65th percentile). A physician rater, without knowledge of hypotheses and predictors, then reviewed the medical records of these patients and identified somatizing patients (n = 122) and nonsomatizing patients (n = 761). In two-thirds of the population (the derivation set), we used logistic regression to refine our hypothesis and identify predictors of somatization available from the MIS: demographic data, all medical encounters, and primary diagnoses made by usual care physicians (ICD-9 codes). We then tested our prediction model in the remaining one-third of the population (the validation set) to validate its usefulness.
The derivation set contained the following significant correlates of somatization: gender, total number of visits, and percent of visits with somatization potential. The c-statistic, equivalent to the area under the ROC curve, was 0.90. In the validation set, the explanatory power was less with a still impressive c-statistic of 0.78. A predicted probability of 0.04 identified almost all somatizers, whereas a predicted probability of 0.40 identified about half of all somatizers but produced few false positives.
We have developed and validated a prediction model from the MIS that helps to distinguish chronic somatizing patients from other high utilizing patients. Our method requires corroboration but carries the promise of providing clinicians and health plan directors with an inexpensive, simple approach for identifying the common somatizing patient and, in turn, targeting them for treatment. The screener does not require clinicians' time.
Medical Care 10/2001; 39(9):968-78. · 2.94 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We describe a method for estimating the marginal cost–effectiveness ratio (CER) of two competing treatments or intervention strategies after adjusting for covariates that may influence the primary endpoint of survival. A Cox regression model is used for modeling covariates and estimates of both the cost and effectiveness parameters, which depend on the survival curve, are obtained from the estimated survival functions for each treatment at a specified covariate. Confidence intervals for the covariate-adjusted CER are presented.
Journal of Statistical Planning and Inference 01/1999; 75(2):291-304. · 0.60 Impact Factor