Marginal indirect standardization using latent clustering on multiple hospitals

Introduction: A number of statistical approaches have been advocated and implemented to estimate adjusted hospital outcomes for public reporting or reimbursement. Nonetheless, the ability of these methods to identify hospital performance outliers in support of quality improvement has not yet been fully investigated. Methods: We leveraged data from patients undergoing coronary artery bypass grafting surgery between 2012-2015 at 33 hospitals participating in a statewide quality collaborative. We applied 5 different statistical approaches (1: indirect standardization with standard logistic regression models, 2: indirect standardization with fixed effect models, 3: indirect standardization with random effect models, 4: direct standardization with fixed effect models, 5: direct standardization with random effect models) to estimate hospital post-operative pneumonia rates adjusting for patients’ risk. Unlike the standard logistic regression models, both fixed effect and random effect models accounted for hospital effect. We applied each method to each year, and subsequently compared methods in their ability to identify hospital performance outliers. Results: Pneumonia rates ranged from 0 % to 24 %. The standard logistic regression models for 2013-2015 had c-statistics of 0.73-0.75, fixed effect models had c-statistics of 0.81-0.83, and random effect models had c-statistics of 0.80-0.83. Each method differed in its ability to identify performance outliers (Figure 1). In direct standardization, random effect models stabilized the hospital rates by moving the estimated rates toward the average rate, fixed effect models produced larger standard errors of hospital effect (particularly for hospitals with low case volumes). In indirect standardization, the three models showed high agreement on their derived observed: expected ratio (intraclass correlation =0.95). Indirect standardization with fixed effect or random effect models, identified similar hospital performance outliers in each year. Conclusion: The five statistical approaches varied in their ability to identify performance outliers. Given its higher sensitivity to outlier hospitals, indirect standardization methods with fixed or random effect models, may be best suited to support quality improvement activities.

Download Full-text

Risk adjustment in maternity care: the use of indirect standardization

International Journal of Women s Health ◽

10.2147/ijwh.s9494 ◽

2010 ◽

pp. 255 ◽

Cited By ~ 1

Author(s):

James Nicholson

Keyword(s):

Risk Adjustment ◽

Maternity Care ◽

Indirect Standardization

Download Full-text

Methodology minute: Indirect standardization

American Journal of Infection Control ◽

10.1016/j.ajic.2017.05.022 ◽

2017 ◽

Vol 45 (9) ◽

pp. 1050-1051

Author(s):

Kathleen M. McMullen

Keyword(s):

Indirect Standardization

Download Full-text

Age Specific and Differential Fertility in Durham and Easington Registration Districts, England, 1851 and 1861

Social Science History ◽

10.1017/s0145553200022288 ◽

1977 ◽

Vol 2 (1) ◽

pp. 23-52 ◽

Cited By ~ 1

Author(s):

Michael R. Haines

Keyword(s):

Fertility Decline ◽

The United States ◽

Extensive Study ◽

Low Fertility ◽

Differential Fertility ◽

Specific Data ◽

Fertility Differentials ◽

Secular Decline ◽

Indirect Standardization ◽

Data Constraints

The current interest in fertility decline and the demographic transition has led to extensive study of the secular decline of fertility in countries and sub-regions of presently low fertility. Extensive work has been completed or is underway regarding Europe in connection with Ansley Coale’s European Fertility Project. In addition, considerable attention has been paid to fertility in the demographic experience of the United States, and to other areas which have experienced fertility decline. One problem with most historical fertility studies is that they lack data on age-specific fertility and also on fertility differentials. So, for example, the European Fertility Project has relied on a form of indirect standardization, the indices of overall fertility (If), marital fertility (Ig), illegitimate fertility (Ih), and proportions married (Im), to compensate for the lack of age-specific data. There are similar historical data constraints on some types of differential fertility categorizations (e.g., social class, literacy, occupation, nativity).

Download Full-text

COVID-19 mortality across selected states in India: the role of age structure.

10.1101/2021.07.18.21259093 ◽

2021 ◽

Author(s):

Mohamed Jainul Azarudeen ◽

Tanzin Dikid ◽

Karishma Kurup ◽

Khyati Aroskar ◽

Himanshu Chauhan ◽

...

Keyword(s):

Mortality Rate ◽

Age Structure ◽

Disease Surveillance ◽

Economic Status ◽

Age Groups ◽

The Elderly ◽

Mortality Rates ◽

Health System Intervention ◽

Indirect Standardization ◽

Standardized Mortality Rate

Background Mortality rates provide an opportunity to identify and act on the health system intervention for preventing deaths. Hence, it is essential to appreciate the influence of age structure while reporting mortality for a better summary of the magnitude of the epidemic. Objectives We described and compared the pattern of COVID-19 mortality standardized by age between selected states and India from January to November 2020. Methods We initially estimated the Indian population for 2020 using the decadal growth rate from the previous census (2011). This was followed by estimations of crude and age-adjusted mortality rate per million for India and the selected states. We used this information to perform indirect standardization and derive the age-standardized mortality rates for the states for comparison. In addition, we derived a ratio for age-standardized mortality to compare across age groups within the state. We extracted information regarding COVID-19 deaths from the Integrated Disease Surveillance Programme special surveillance portal up to November 16, 2020. Results The crude mortality rate of India stands at 88.9 per million population(118,883/1,337,328,910). Age-adjusted mortality rate (per million) was highest for Delhi (300.5) and lowest for Kerala (35.9).The age-standardized mortality rate (per million) for India is (<15 years=1.6, 15-29 years=6.3, 30-44 years=35.9, 45-59 years=198.8, 60-74 years=571.2, & ≥75 years=931.6). The ratios for age-standardized mortality increase proportionately from 45-59 years age group across all the states. Conclusion There is high COVID-19 mortality not only among the elderly ages, but we also identified heavy impact of COVID-19 on the working population. Therefore, we recommend further evaluation of age-adjusted mortality for all States and inclusion of variables like gender, socio-economic status for standardization while identifying at-risk populations and implementing priority public health actions. Keywords COVID-19, Mortality, Age Standardized Mortality Rate, Indirect Standardization.

Download Full-text

Indirect Standardization of Neonatal Mortality for Birth Weight

International Journal of Epidemiology ◽

10.1093/ije/11.2.146 ◽

1982 ◽

Vol 11 (2) ◽

pp. 146-154 ◽

Cited By ~ 7

Author(s):

JOEL C KLEINMAN

Keyword(s):

Birth Weight ◽

Neonatal Mortality ◽

Indirect Standardization

Download Full-text

The risk of cancer among Lithuanian medical radiation workers in 1978–2004

Medicina ◽

10.3390/medicina45050052 ◽

2009 ◽

Vol 45 (5) ◽

pp. 412

Author(s):

Vitalija Samerdokienė ◽

Vydmantas Atkočius ◽

Konstantinas Valuckas

Keyword(s):

Cancer Risk ◽

Risk Evaluation ◽

Incidence Rates ◽

Medical Radiation ◽

Radiation Workers ◽

Indirect Standardization ◽

And Gender ◽

Risk Of Cancer ◽

Gender Specific

Objectives. To describe the cohort of Lithuanian medical radiation workers and to estimate the risk of cancer during 1978–2004. Methods. Analysis of cancer risk evaluation was done using the retrospective cohort method, an indirect standardization. The observed numbers of cancer cases were obtained from the National Cancer Registry. The expected numbers were based on the age- and gender-specific incidence rates for the general Lithuanian population in 5-year periods. The standardized incidence ratios and 95% confidence intervals (assuming that the incidence of cancer follows the Poisson distribution) were calculated. Results. During the follow-up of 1978–2004, 159 cases of cancer were observed. There was no increased overall cancer risk in men (SIR=0.92, 95% CI=0.62–1.33, based on 29 cases) and women (SIR=0.97, 95% CI=0.81–1.15, based on 130 cases). The risk of leukemia among men and women was insignificantly increased. Conclusions. During the follow-up period, the overall cancer risk among medical radiation workers was the same as in the general population of Lithuania

Download Full-text

The Non-Decline in U.S. Twin Birth Rates, 1964–1983

Acta geneticae medicae et gemellologiae twin research ◽

10.1017/s0001566000006061 ◽

1987 ◽

Vol 36 (3) ◽

pp. 313-323 ◽

Cited By ~ 12

Author(s):

Gordon Allen

Keyword(s):

United States ◽

Birth Order ◽

Maternal Age ◽

The United States ◽

Birth Rates ◽

Continuous Increase ◽

Twin Birth ◽

Indirect Standardization ◽

Adjusted Rate ◽

Nearly Continuous

AbstractDetailed twin birth rates for the United States are unavailable since 1964. In 1983 the crude twinning rate for women of white race was higher than in 1964, but there had been great changes in maternal age and parity. Indirect standardization for maternal age and birth order provides estimated total twinning rates that can be compared over the entire period. The adjusted rates for whites show a nearly continuous increase except after a 2-year reporting hiatus, 1969-70, when rates dropped back 10%. In blacks the adjusted rate increased between 1966 and 1978, except for the 1968-71 shift. The distributions of rate increases by maternal age and by race argue against effects of medical ovulation stimulants, but a disproportionate increase of triplets argues for such effects. Study is needed of rates specific for maternal age and parity, rather than of total rates.

Download Full-text