Disparate Estimates Research Articles

Standardized data collection practices and the racial/ethnic distribution of hospitalized patients.

Although frequently used to track health care disparities, patient race/ethnicity data collected by hospitals can be unreliable, particularly for smaller minority groups. We sought to determine whether the racial/ethnic distribution of hospitalized patients shifted after implementation of a statewide initiative to standardize data collection practices. We conducted a difference-in-differences analysis of the State Inpatient Databases to estimate changes in the proportion of patients identified as non-Hispanic white, non-Hispanic black, Hispanic, Asian/Pacific Islander, and "other," before (2005-2006) and after (2008-2009) standardized practices were implemented in New Jersey relative to New York, a state with similar demographics but no changes to data collection. Among 12,552,702 hospital discharges, modest relative changes were noted in the proportion of patients identified as non-Hispanic white [+1.1%; 95% confidence interval (CI): +0.9 to +1.2] and non-Hispanic black (+1.6%; 95% CI: +1.1 to +2.1) in New Jersey that were attributed to its use of standardized data collection practices as compared with New York. Larger relative changes were noted in the proportion of patients identified as Hispanic (-7.1%; 95% CI: -7.8 to -6.4), Asian/Pacific Islander (+26.5%; 95% CI: +25.1 to +27.9), and "other" (-24.6%; 95% CI: -26.4 to -22.8). This pattern was largely consistent in analyses stratified by sex, age, and Major Diagnostic Category. Measurement of health care disparities fundamentally depends on the racial/ethnic categorization of individuals. By redistributing substantial proportions of patients across smaller minority groups, standardized data collection could lead to shifts in estimates of health care disparities for these rapidly growing populations.

Does Extending Unemployment Benefits Improve Job Quality?

Contrary to standard search models predictions, past studies have not found a positive effect of unemployment insurance (UI) on reemployment wages. We estimate a positive UI wage effect exploiting an age-based regression discontinuity design in Austria. A search model incorporating duration dependence predicts two countervailing forces: UI induces workers to seek higher-wage jobs, but reduces wages by lengthening unemployment. Matching-function heterogeneity plausibly generates a negative relationship between the UI unemployment-duration and wage effects, which holds empirically in our sample and across studies, reconciling disparate wage-effect estimates. Empirically, UI raises wages by improving reemployment firm quality and attenuating wage drops. (JEL J31, J64, J65)

All else being equal, men and women are still not the same: using risk models to understand gender disparities in care.

Despite growing awareness of the burden of cardiovascular disease (CVD) in women over the past 2 decades, reports of disparities in the delivery of health care for these diseases have persisted over time.1 Under-treatment of women as compared with men has been described for primary prevention, stable coronary artery disease (CAD) and suspected or diagnosed acute coronary syndromes in both observational studies and controlled experiments.2–6 However, because men and women are nonidentical in many ways that might be relevant for treatment decisions, a critical question for policy makers and clinicians is whether these gender-based variations in treatment correspond to lower quality care. A demonstration of clinical inequity (unequal treatment despite equal clinical need) rather than strict inequality (unequal treatment regardless of need or condition) has been proposed as a framework to identify inappropriate variations in healthcare utilization.7 Following this model, it is important to consider whether observed gender differences correspond to inequities or may reflect appropriate variations explained by differences in treatment indications and contraindications or patient preferences.7 In a well-known study published over a decade ago, investigators devised a clever experiment to address just this question of identifying inappropriate healthcare variations.8 Using actors posing as patients with chest pain, primary care physicians were less likely to refer women for cardiac catheterization than men. Because these patients were carefully designed to be identical in age, cardiac risk factors, and symptoms, followed a script standardizing patient communication, and only differed by gender (and race), it seemed obvious that biased decision-making must account for treatment differences. In this commentary, we illustrate why analyses that attempt to control for all known cardiac risk factors that affect treatment decisions might still yield misleading findings about the presence, magnitude, and causes of gender disparities. This residual bias occurs …

Does Extending Unemployment Benefits Improve Job Quality?

Contrary to standard search model predictions, prior studies failed to estimate a positive effect of unemployment insurance (UI) on reemployment wages. This paper estimates a positive UI wage effect exploiting an age-based regression discontinuity in Austrian administrative data. A search model incorporating duration dependence determines the UI wage effect as the balance between two offsetting forces: UI causes agents to seek higher-wage jobs, but also reduces wages by lengthening unemployment. This implies a negative relationship between the UI unemployment duration and wage effects, which holds empirically both in our sample and across studies, reconciling disparate wage-effect estimates. Empirically, UI raises wages by improving reemployment firms' quality and attenuating wage drops.

An evaluation of fossil tip-dating versus node-age calibrations in tetraodontiform fishes (Teleostei: Percomorphaceae)

Time-calibrated phylogenies based on molecular data provide a framework for comparative studies. Calibration methods to combine fossil information with molecular phylogenies are, however, under active development, often generating disagreement about the best way to incorporate paleontological data into these analyses. This study provides an empirical comparison of the most widely used approach based on node-dating priors for relaxed clocks implemented in the programs BEAST and MrBayes, with two recently proposed improvements: one using a new fossilized birth–death process model for node dating (implemented in the program DPPDiv), and the other using a total-evidence or tip-dating method (implemented in MrBayes and BEAST). These methods are applied herein to tetraodontiform fishes, a diverse group of living and extinct taxa that features one of the most extensive fossil records among teleosts. Previous estimates of time-calibrated phylogenies of tetraodontiforms using node-dating methods reported disparate estimates for their age of origin, ranging from the late Jurassic to the early Paleocene (ca. 150–59Ma). We analyzed a comprehensive dataset with 16 loci and 210 morphological characters, including 131 taxa (95 extant and 36 fossil species) representing all families of fossil and extant tetraodontiforms, under different molecular clock calibration approaches. Results from node-dating methods produced consistently younger ages than the tip-dating approaches. The older ages inferred by tip dating imply an unlikely early-late Jurassic (ca. 185–119Ma) origin for this order and the existence of extended ghost lineages in their fossil record. Node-based methods, by contrast, produce time estimates that are more consistent with the stratigraphic record, suggesting a late Cretaceous (ca. 86–96Ma) origin. We show that the precision of clade age estimates using tip dating increases with the number of fossils analyzed and with the proximity of fossil taxa to the node under assessment. This study suggests that current implementations of tip dating may overestimate ages of divergence in calibrated phylogenies. It also provides a comprehensive phylogenetic framework for tetraodontiform systematics and future comparative studies.

Outcomes and unmet need for neonatal surgery in a resource-limited environment: Estimates of global health disparities from Kampala, Uganda

PurposeReported outcomes of neonatal surgery in low-income countries (LICs) are poor. We examined epidemiology, outcomes, and met and unmet need of neonatal surgical diseases in Uganda. MethodsPediatric general surgical admissions and consults from January 1, 2012, to December 31, 2012, at a national referral center in Uganda were analyzed using a prospective database. Outcomes were compared with high-income countries (HICs), and met and unmet need was estimated using burden of disease metrics (disability-adjusted life years or DALYs). Results23% (167/724) of patients were neonates, and 68% of these survived. Median age of presentation was 5days, and 53% underwent surgery. 88% survived postoperatively, while 55% died without surgery (p<0.001). Gastroschisis carried the highest mortality (100%) and the greatest mortality disparity with HICs. An estimated 5072 DALYs were averted by neonatal surgery in Uganda (met need), with 140,154 potentially avertable (unmet need). Approximately 3.5% of the need for neonatal surgery is met by the health system. ConclusionsMore than two thirds of surgical neonates survived despite late presentation and lack of critical care. Epidemiology and outcomes differ greatly with HICs. A high burden of hidden mortality exists, and only a negligible fraction of the population need for neonatal surgery is met by health services.

Significance of HbA1c and its measurement in the diagnosis of diabetes mellitus: US experience.

The 2014 American Diabetes Association guidelines denote four means of diagnosing diabetes. The first of these is a glycosylated hemoglobin (HbA1c) >6.5%. This literature review summarizes studies (n=47) in the USA examining the significance, strengths, and limitations of using HbA1c as a diagnostic tool for diabetes, relative to other available means. Due to the relatively recent adoption of HbA1c as a diabetes mellitus diagnostic tool, a hybrid systematic, truncated review of the literature was implemented. Based on these studies, we conclude that HbA1c screening for diabetes has been found to be convenient and effective in diagnosing diabetes. HbA1c screening is particularly helpful in community-based and acute care settings where tests requiring fasting are not practical. Using HbA1c to diagnose diabetes also has some limitations. For instance, HbA1c testing may underestimate the prevalence of diabetes, particularly among whites. Because this bias differs by racial group, prevalence and resulting estimates of health disparities based on HbA1c screening differ from those based on other methods of diagnosis. In addition, existing evidence suggests that HbA1c screening may not be valid in certain subgroups, such as children, women with gestational diabetes, patients with human immunodeficiency virus, and those with prediabetes. Further guidelines are needed to clarify the appropriate use of HbA1c screening in these populations.

The importance of historical data in simulating malaria epidemiology and control interventions: application to multiple sites in Madagascar

Planning malaria interventions requires the prediction of likely impacts of different intervention strategies. Simulation models can provide such predictions but weak information about pre-control levels of transmission, intervention coverage and access to care often makes it challenging to correctly parameterize them. We consider a number of low malaria transmission sites in Madagascar, with available historical prevalence and entomological inoculation rate (EIR) estimates (by mosquito sampling), giving disparate estimates of historical exposure. Information about implementation of Long Lasting Insecticide Nets (LLINs) and Indoor Residual Spraying (IRS), and access to healthcare, collated from malaria surveys and on-going cross-sectional studies were used to parameterise simulations of malaria transmission, prevalence and burden within the OpenMalaria platform. Multiple parameterisations were considered using various sources of data for pre-intervention transmission level, intervention coverage and access to healthcare. In some sites the simulated impact of existing vector control programs matches reasonably well the malaria prevalence measured in a recent national survey. In others it predicts lower than observed prevalence, very likely because the models do not capture residual local transmission foci. The simulations suggest that the most cost-effective vector control strategy would be to scale-up LLINs or IRS only, depending on the transmission level. Indeed, preliminary results show no additional benefit of IRS where LLINs were used. These preliminary results suggest that historical prevalence data, combined with current coverage information are potentially adequate for planning intervention strategies. The outcome of intervention scale-up is essentially unpredictable if baseline information is poor. Reproducing the observed epidemiology of malaria through simulations both provides confidence in the use of the model but serves as a basis for prospective studies that support decision-making, including cost-effectiveness analyses.

Ethnic Identification and its Consequences for Measuring Inequality in Mexico.

This paper examines ethnic boundary crossing and its impact on estimates of ethnic disparities in children's outcomes in the specific context of Mexico, a country with the largest indigenous population in the Western hemisphere. The boundary that separates the indigenous and non-indigenous population is known to be extremely fluid as it is based on characteristics that can easily change within a generation such as language use, cultural practices and a subjective sense of belonging. Using data from the Mexican census I examine the ethnic classification of children of indigenous parents. I find that movement across the ethnic boundary depends on which of the two criteria currently recognized by the Mexican census is used. Children of indigenous parents are much less likely to be classified as indigenous according to language proficiency, especially when their parents have higher levels of education. By contrast, when proxy self-identification is used as a criterion, children of indigenous parents are more likely to be classified as indigenous, and greater parental education actually results in higher odds that children will be classified as indigenous. The shift in children's indigenous classification with parental education is found to strongly affect estimates of educational disparities between indigenous and non-indigenous children.

Ethnicity in England: What Parents' Country of Birth Can and Can't Tell Us about Their Children's Ethnic Identification

Despite the importance of adequately measuring ethnicity to keep track of ethnic disparities in important outcomes, there is little consensus on how this is best achieved. Different countries apply widely differing standards, seemingly guided by convenience more than conviction. We employ unique data on 3035 15-year-olds living in England to investigate inconsistencies when two different measures are used: (i) ethnic group self-identification by respondents and (ii) information on parents' birth country. The former is currently standard in England and Wales and the latter in many European countries. After having aligned our country-of-origin-based measure to the ethnic categories currently in use in England and Wales, we are able to show substantial agreement across measures, and estimates of disparities in language proficiency and delinquency appear little affected by the choice of measure. We conclude that comparative research using secondary data need not be badly biased due to inconsistent measures if sufficient care is taken in harmonisation.

Accuracy of parental reports of children's HPV vaccine status: implications for estimates of disparities, 2009-2010.

Since the introduction of the human papillomavirus (HPV) vaccine in 2006, there have been considerable efforts at the national and state levels to monitor uptake and better understand the individual and system-level factors that predict who gets vaccinated. A common method of measuring the vaccination status of adolescents is through parental recall. We examined how the accuracy of parents' reports of their daughters' HPV vaccination status varied by social characteristics. Data were taken from the 2009-2010 National Immunization Survey (NIS)-Teen, which includes a household interview and a provider-completed immunization history. We evaluated concordance between parents' and providers' reports of teens' HPV vaccine initiation (≥1 dose) and completion (≥3 doses). We assessed bivariate associations of sociodemographic characteristics with having a concordant, false-positive (overreporting) or false-negative (underreporting) report, and used multinomial logistic regression to estimate the independent impact of each characteristic. In bivariate analyses, concordance of parent-reported HPV vaccine initiation was associated with each of the sociodemographic characteristics investigated. In regression models, self-reported nonwhite race, lower household income, and lower education level of the teen's mother were associated with a higher likelihood of having a false-negative parental report than a concordant report. Our results indicate that, while estimates of overall coverage based on parental report may be unbiased, the differences in the accuracy of parental report could result in misleading estimates of disparities in HPV vaccine coverage.

Operational Definitions of Sexual Orientation and Estimates of Adolescent Health Risk Behaviors.

Increasing attention to the health of lesbian, gay, and bisexual (LGB) populations comes with requisite circumspection about measuring sexual orientation in surveys. However, operationalizing these variables also requires considerable thought. This research sought to document the consequences of different operational definitions of sexual orientation by examining variation in health risk behaviors. Using Massachusetts Youth Risk Behavior Survey data, we examined how operational definitions of sexual behavior and sexual identity influenced differences among three health behaviors known to disparately affect LGB populations: smoking, suicide risk, and methamphetamine use. Sexual behavior and sexual identity were also examined together to explore if they captured unique sources of variability in behavior. Estimates of health disparities changed as a result of using either sexual behavior or sexual identity. Youth who reported their sexual identity as "not sure" also had increased odds of health risk behavior. Disaggregating bisexual identity and behavior from same-sex identity and behavior frequently resulted in the attenuation or elimination of health disparities that would have otherwise been attributable to exclusively same-sex sexual minorities. Finally, sexual behavior and sexual identity explained unique and significant sources of variability in all three health behaviors. Researchers using different operational definitions of sexual orientation could draw different conclusions, even when analyzing the same data, depending upon how they chose to represent sexual orientation in analyses. We discuss implications that these manipulations have on data interpretation and provide specific recommendations for best-practices when analyzing sexual orientation data collected from adolescent populations.

Measuring disparities in sanitation access: does the measure matter?

Initiatives to monitor progress in health interventions like sanitation are increasingly focused on disparities in access. We explored three methodological challenges to monitoring changes in sanitation coverage across socio-economic and demographic determinants: (i) confounding by wealth indices including water and sanitation assets, (ii) use of individual urban and rural settings versus national wealth indices and (iii) child-level versus household-level analyses. Sanitation coverage by wealth for children and households across settings was estimated from recent Demographic and Health Surveys in six low-income countries. Household assignment to wealth quintiles was based on principal components analyses of assets. Concordance in household quintile assignment and estimated distribution of improved sanitation was assessed using two wealth indices differing by inclusion or exclusion of water and sanitation assets and independently derived for each setting. Improved sanitation was estimated using under five children and households. Wealth indices estimated with water, and sanitation assets are highly correlated with indices excluding them but can overstate disparities in sanitation access. Independently, derived setting wealth indices highly correlate with setting estimates of coverage using a single national index. Sanitation coverage and quintile disparities were consistently lower in household-level estimates. Standard asset indices provide a reasonably robust measure of disparities in improved sanitation, although overestimation is possible. Separate setting wealth quintiles reveal important disparities in urban areas, but analysis of setting quintiles using a national index is sufficient. Estimates and disparities in household-level coverage of improved sanitation can underestimate coverage for children under five.

Statistical methods to enhance reporting of Aboriginal Australians in routine hospital records using data linkage affect estimates of health disparities.

To investigate under-recording of Aboriginal people in hospital data from New South Wales (NSW), Australia, define algorithms for enhanced reporting, and examine the impact of these algorithms on estimated disparities in cardiovascular and injury outcomes. NSW Admitted Patient Data were linked with NSW mortality data (2001-2007). Associations with recording of Aboriginal status were investigated using multilevel logistic regression. The number of admissions reported as Aboriginal according to six algorithms was compared with the original (unenhanced) Aboriginal status variable. Age-standardised admission, and 30- and 365-day mortality ratios were estimated for cardiovascular disease and injury. Sixty per cent of the variation in recording of Aboriginal status was due to the hospital of admission, with poorer recording in private and major city hospitals. All enhancement algorithms increased the number of admissions reported as Aboriginal, from between 4.1% and 37.8%. Admission and mortality ratios varied markedly between algorithms, with less strict algorithms resulting in higher admission rate ratios, but generally lower mortality rate ratios, particularly for cardiovascular disease. The choice of enhancement algorithm has an impact on the number of people reported as Aboriginal and on estimated outcome ratios. The influence of the hospital on recording of Aboriginal status highlights the importance of continued efforts to improve data collection. Estimates of Aboriginal health disparity can change depending on how Aboriginal status is reported. Sensitivity analyses using a number of algorithms are recommended.

The Time for Austerity: Estimating the Average Treatment Effect of Fiscal Policy

Elevated government debt levels in advanced economies have risen rapidly as sovereigns absorbed private sector losses and cyclical deficits blew up in the Global Financial Crisis and subsequent slump. A rush to fiscal austerity followed but its justifications and impacts have been heavily debated. Research on the effects of austerity on macroeconomic aggregates remains unsettled, mired by the difficulty of identifying multipliers from observational data. This paper reconciles seemingly disparate estimates of multipliers within a unified framework. We do this by first evaluating the validity of common identification assumptions used by the literature and find that they are largely violated in the data. Next, we use new propensity score methods for time-series data with local projections to quantify how contractionary austerity really is, especially in economies operating below potential. We find that the adverse effects of austerity may have been understated.

Reconciling disparate estimates of total offset on the southern San Andreas fault

Late Cenozoic tectonic reconstructions of southern California (United States) have been hindered for decades by disagreement over the total displacement on the southern San Andreas fault; disparate estimates of dextral offset range from 160 km to 240 km. Prior estimates were based on purely translational models that neglect the effects of transrotational strain on the orientation and distribution of key geologic markers. We present a new reconstruction that integrates published lithologic, structural, and paleocurrent data with the effects of clockwise rotation in the eastern Transverse Ranges (ETR). This model yields 200 ± 14 km of cumulative offset on the southern San Andreas fault system in the Salton Trough. The disparity between our estimate and a lower estimate of 160 km for the Mojave segment can be explained by a combination of transtensional elongation along the southwest boundary of the ETR, transfer of strain to the eastern California shear zone, and diffuse crustal shortening in the Big Bend region. This reconstruction reconciles previously incompatible and long-debated models, and places an important new constraint on the tectonic evolution of the Pacific–North America plate boundary in southern California.

Accounting for Clinical Action Reduces Estimates of Gender Disparities in Lipid Management for Diabetic Veterans

ABSTRACTBACKGROUNDWomen with diabetes have higher low-density lipoprotein (LDL) levels than men, resulting in apparent disparities between genders on quality indicators tied to LDL thresholds.OBJECTIVETo investigate whether gender disparities persist when accounting for clinical action with statins or cardiovascular risk.DESIGNRetrospective cohort study.PARTICIPANTSVeterans Health Administration patients (21,780 women and 646,429 men) aged 50–75 with diabetes.MAIN MEASURESThreshold measure: LDL < 100 mg/dL; clinical action measure: LDL < 100 mg/dL; or LDL ≥ 100 mg/dL and the patient was prescribed a moderate or high-dose statin at the time of the test; or LDL ≥ 100 mg/dL and the patient received other appropriate clinical action within 90 days; adherence: continuous multiple interval measure of gaps in dispensed medication (CMG).KEY RESULTSWomen were much less likely to have LDL < 100 mg/dL than were men (55 % vs. 68 %). This disparity narrowed from 13 % to 6 % for passing the clinical action measure (79 % vs. 85 %). These gender differences persisted among those with ischemic heart disease (IHD). Women had a lower odds of passing the clinical action measure (odds ratio 0.68, 95 % confidence interval 0.66–0.71). Among those with IHD, the gender gap increased with age. Differences in pass rates were explained by women’s higher LDL levels, but not by their slightly worse adherence (3 % higher CMG).CONCLUSIONSWomen and men veterans receive more similar quality of care for lipids in diabetes than previously indicated. Less reassuringly, the remaining gender differences appear to be as common in women at high cardiovascular risk as in those at low risk. Rather than focus on simply improving LDL levels in all women with diabetes, future efforts should ensure that patients with high cardiovascular risk are appropriately treated with statins when clinically indicated, feasible, and concordant with patient preferences.

Differences in Diffusion of FDA Antidepressant Risk Warnings Across Racial-Ethnic Groups

OBJECTIVE Numerous articles have identified that medical technologies diffuse more rapidly among non-Latino whites compared with other racial-ethnic groups. However, whether health risk warnings also diffuse differentially across racial-ethnic minority groups is uncertain. This study assessed racial-ethnic variation in children's antidepressant use before and after the 2004 black-box warning concerning risks of antidepressants for youths. METHODS Data consisted of responses for white, black, and Latino youths ages five through 17 from the 2002-2008 Medical Expenditure Panel Survey (N=44,422). The dependent variable was any antidepressant use in the prior year. Independent variables were race-ethnicity, year, psychological impairment, income, insurance status, region, and parents' education level. Logistic regression models were used to assess antidepressant use conditional on race-ethnicity, time, interaction between race-ethnicity and time, need, socioeconomic status, and Institute of Medicine-concordant estimates of disparities in predicted antidepressant use before and after the warning. RESULTS The warnings affected antidepressant use differentially for whites, blacks, and Latinos. Usage rates among whites decreased from 3.3 to 2.1 percentage points between prewarning and postwarning, whereas usage rates remained steady among Latinos and increased among blacks. Findings were significant in multiple regression analyses, in which predictions were adjusted for need. CONCLUSIONS The findings indicate that health safety information on antidepressant usage among children diffused faster among whites than nonwhites, suggesting the need to improve infrastructure for delivering important health messages to racial-ethnic minority populations.

Benchmarking the Incidence and Mortality of Severe Sepsis in the United States*

In 1992, the first consensus definition of severe sepsis was published. Subsequent epidemiologic estimates were collected using administrative data, but ongoing discrepancies in the definition of severe sepsis produced large differences in estimates. We seek to describe the variations in incidence and mortality of severe sepsis in the United States using four methods of database abstraction. We hypothesized that different methodologies of capturing cases of severe sepsis would result in disparate estimates of incidence and mortality. Using a nationally representative sample, four previously published methods (Angus et al, Martin et al, Dombrovskiy et al, and Wang et al) were used to gather cases of severe sepsis over a 6-year period (2004-2009). In addition, the use of new International Statistical Classification of Diseases, 9th Edition (ICD-9), sepsis codes was compared with previous methods. Annual national incidence and in-hospital mortality of severe sepsis. The average annual incidence varied by as much as 3.5-fold depending on method used and ranged from 894,013 (300/100,000 population) to 3,110,630 (1,031/100,000) using the methods of Dombrovskiy et al and Wang et al, respectively. Average annual increase in the incidence of severe sepsis was similar (13.0% to 13.3%) across all methods. In-hospital mortality ranged from 14.7% to 29.9% using abstraction methods of Wang et al and Dombrovskiy et al. Using all methods, there was a decrease in in-hospital mortality across the 6-year period (35.2% to 25.6% [Dombrovskiy et al] and 17.8% to 12.1% [Wang et al]). Use of ICD-9 sepsis codes more than doubled over the 6-year period (158,722 - 489,632 [995.92 severe sepsis], 131,719 - 303,615 [785.52 septic shock]). There is substantial variability in incidence and mortality of severe sepsis depending on the method of database abstraction used. A uniform, consistent method is needed for use in national registries to facilitate accurate assessment of clinical interventions and outcome comparisons between hospitals and regions.

Sight and sound out of synch: Fragmentation and renormalisation of audiovisual integration and subjective timing

The sight and sound of a person speaking or a ball bouncing may seem simultaneous, but their corresponding neural signals are spread out over time as they arrive at different multisensory brain sites. How subjective timing relates to such neural timing remains a fundamental neuroscientific and philosophical puzzle. A dominant assumption is that temporal coherence is achieved by sensory resynchronisation or recalibration across asynchronous brain events. This assumption is easily confirmed by estimating subjective audiovisual timing for groups of subjects, which is on average similar across different measures and stimuli, and approximately veridical. But few studies have examined normal and pathological individual differences in such measures.Case PH, with lesions in pons and basal ganglia, hears people speak before seeing their lips move. Temporal order judgements (TOJs) confirmed this: voices had to lag lip-movements (by ∼200 msec) to seem synchronous to PH. Curiously, voices had to lead lips (also by ∼200 msec) to maximise the McGurk illusion (a measure of audiovisual speech integration). On average across these measures, PH's timing was therefore still veridical. Age-matched control participants showed similar discrepancies. Indeed, normal individual differences in TOJ and McGurk timing correlated negatively: subjects needing an auditory lag for subjective simultaneity needed an auditory lead for maximal McGurk, and vice versa. This generalised to the Stream–Bounce illusion. Such surprising antagonism seems opposed to good sensory resynchronisation, yet average timing across tasks was still near-veridical.Our findings reveal remarkable disunity of audiovisual timing within and between subjects. To explain this we propose that the timing of audiovisual signals within different brain mechanisms is perceived relative to the average timing across mechanisms. Such renormalisation fully explains the curious antagonistic relationship between disparate timing estimates in PH and healthy participants, and how they can still perceive the timing of external events correctly, on average.