Proportional Hazards Regression in Epidemiologic Follow-up Studies: An Intuitive Consideration of Primary Time Scale
The DOI matches its Crossref record; 43 of the 43 references it cites were found in a registry or an index.
Abstract
In epidemiologic cohort studies of chronic diseases, such as heart disease or cancer, confounding by age can bias the estimated effects of risk factors under study. With Cox proportional-hazards regression modeling in such studies, it would generally be recommended that chronological age be handled nonparametrically as the primary time scale. However, studies involving baseline measurements of biomarkers or other factors frequently use follow-up time since measurement as the primary time scale, with no explicit justification. The effects of age are adjusted for by modeling age at entry as a parametric covariate. Parametric adjustment raises the question of model adequacy, in that it assumes a known functional relationship between age and disease, whereas using age as the primary time scale does not. We illustrate this graphically and show intuitively why the parametric approach to age adjustment using follow-up time as the primary time scale provides a poor approximation to age-specific incidence. Adequate parametric adjustment for age could require extensive modeling, which is wasteful, given the simplicity of using age as the primary time scale. Furthermore, the underlying hazard with follow-up time based on arbitrary timing of study initiation may have no inherent meaning in terms of risk. Given the potential for biased risk estimates, age should be considered as the preferred time scale for proportional-hazards regression with epidemiologic follow-up data when confounding by age is a concern.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Multivariable adjusted estimates of the effect of quintiles of step count on the relative hazards of cardiovascular mortality and all-cause mortality were derived using Cox proportional hazards regression using age as the underlying timescale30,31.”
From Development and Validation of a Machine Learning Wrist-worn Step Detection Algorithm with Deployment in the UK Biobank (2023) · cites this paper as [31]
Citations
Cites
Regression Models and Life-Tables
Cox, D. R. (1972) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1972.tb00899.x · Cited by 39,793
Regression Models and Life-Tables
Cox, D. R. (1992) · Springer Series in Statistics · doi:10.1007/978-1-4612-4380-9_37 · Cited by 26,297
Robust Locally Weighted Regression and Smoothing Scatterplots
Cleveland, W. S. (1979) · Journal of the American Statistical Association · doi:10.1080/01621459.1979.10481038 · Cited by 11,594 · cited 2×
“Data were fit using proportional hazards regression with the Peanuts module in the Epicure software.31 Baseline hazards were smoothed in S-plus (v. 6.2 for Windows, Insightful Corp., Seattle WA) using Lowess.36 Cancer surveillance was performed through the end of 1999 (total 35 years of follow-up).”
MethodInfluentialThe Statistical Analysis of Failure Time Data.
Aitkin, M., Kalbfleisch, J. D., & Prentice, R. L. (1981) · Biometrics · doi:10.2307/2530581 · Cited by 9,877
Modeling Survival Data: Extending the Cox Model
Therneau, T. M., & Grambsch, P. M. (2000) · Statistics for Biology and Health · doi:10.1007/978-1-4757-3294-8 · Cited by 8,674
Cox, D. R., & Oakes, D. (2018) · Chapman & Hall/CRC eBooks · doi:10.1201/9781315137438 · Cited by 6,234
Statistical Methods in Cancer Research. Volume II – The Design and Analysis of Cohort Studies
Juul, S. (1989) · Scandinavian Journal of Social Medicine · doi:10.1177/140349488901700214 · Cited by 3,592
Statistical Models Based on Counting Processes
Bry, X., Andersen, P. K., Borgan, Ø., Gill, R. D., Keiding, N., & Borgan, O. (1997) · Population (French Edition) · doi:10.2307/1534625 · Cited by 2,884
Regression Using Fractional Polynomials of Continuous Covariates: Parsimonious Parametric Modelling
Royston, P., & Altman, D. G. (1994) · Applied Statistics · doi:10.2307/2986270 · Cited by 1,798 · cited 1×
“portion of the proportional hazards model using categories, regression splines,(40,41) or fractional polynomials.(42) With Poisson regression based on person-time rate denominators FIGURE 4.”
MethodHulley, S. B. (1997) · Cited by 1,753
Solid Cancer Incidence in Atomic Bomb Survivors: 1958–1998
Preston, D. L., Ron, E., Tokuoka, S., Funamoto, S., Nishi, N., Soda, M., Mabuchi, K., & Kodama, K. (2007) · Radiation Research · doi:10.1667/rr0763.1 · Cited by 1,753
A case-cohort design for epidemiologic cohort studies and disease prevention trials
PRENTICE, R. L. (1986) · Biometrika · doi:10.1093/biomet/73.1.1 · Cited by 1,439 · cited 1×
“risks being estimated and potential confounders to be adjusted.(12,25,38) Many applications of proportional hazards re-”
BackgroundMethods in Observational Epidemiology
Kelsey, J. L. (1986) · Cited by 1,307
Statistical Methods in Epidemiology
(2012) · SpringerReference · doi:10.1007/springerreference_205629 · Cited by 1,124
Logistic regression, survival analysis, and the Kaplan-Meier curve.
Efron, B. (1988) · Journal of the American Statistical Association · doi:10.1080/01621459.1988.10478612 · Cited by 1,065
Time-to-Event Analysis of Longitudinal Follow-up of a Survey: Choice of the Time-scale
Korn, E. L., Graubard, B. I., & Midthune, D. (1997) · American Journal of Epidemiology · doi:10.1093/oxfordjournals.aje.a009034 · Cited by 1,007 · cited 1×
“Unless all study subjects are the same age at the start of follow-up, the only situation in which the follow-up-time approach can even qualitatively account for the effects of age is when the distribution of event ages is exponential.(8) However, with most diseases, it may be difficult to establish that age at onset i…”
BackgroundRegression Models in Clinical Studies: Determining Relationships Between Predictors and Response
Harre, F. E., Lee, K. L., & Pollock, B. G. (1988) · JNCI Journal of the National Cancer Institute · doi:10.1093/jnci/80.15.1198 · Cited by 957
Research Methods in Occupational Epidemiology
Checkoway, H., Pearce, N. E., & Kriebel, D. (2004) · Oxford University Press eBooks · doi:10.1093/acprof:oso/9780195092424.001.0001 · Cited by 892
Semiparametric Theory and Missing Data
(2006) · Springer Series in Statistics · doi:10.1007/0-387-37345-4 · Cited by 827 · cited 1×
“In fact, 0(t) is often considered to be a nuisance parameter because it can be ignored when estimating .(29) Multiplying by r (z) raises or lowers the hazard (t;z), at time t proportionately relative to the baseline hazard 0(t), according to a specific risk factor profile z.”
BackgroundChoice of time‐scale in Cox's model analysis of epidemiologic cohort data: a simulation study
Thiébaut, A. C. M., & Bénichou, J. (2004) · Statistics in Medicine · doi:10.1002/sim.2098 · Cited by 627 · cited 1×
“When there is confounding by age, that approach fails to remove confounding bias.(13) Although we did not perform a systematic review, examples of analyses based on this approach were easy to find.”
BackgroundMultiplicative Models and Cohort Analysis
Breslow, N. E., Lubin, J. H., Marek, P., & Langholz, B. (1983) · Journal of the American Statistical Association · doi:10.1080/01621459.1983.10477915 · Cited by 565
Noncancer Disease Incidence in Atomic Bomb Survivors, 1958–1998
Yamada, M., Lennie Wong, F., Fujiwara, S., Akahoshi, M., & Suzuki, G. (2004) · Radiation Research · doi:10.1667/rr3183 · Cited by 344 · cited 1×
“ILLUSTRATION We used data on solid cancer incidence among atomic bomb survivors,(34) restricted to those who are participants in the Adult Health Study clinical cohort.(35) The data are part of a study of effects of inflammation on risk of cancer from radiation and smoking (Hsu WL, Neriishi K, Cologne JB, Tatsukawa Y,…”
MethodSelf-citationEpidemiologic Methods: Studying the Occurrence of Illness
Koepsell, T. D., & Weiss, N. S. (2014) · Cited by 258
Critical appraisal of epidemiological studies and clinical trials
Elwood, M. (1998) · Cited by 220
Lamarca, R., Alonso, J., Gomez, G., & Munoz, A. (1998) · The Journals of Gerontology Series A: Biological Sciences and Medical Sciences · doi:10.1093/gerona/53a.5.m337 · Cited by 212
General Relative-Risk Models for Survival Time and Matched Case-Control Analysis
Thomas, D. C. (1981) · Biometrics · doi:10.2307/2530149 · Cited by 183 · cited 1×
“Alternative models include additive hazards(26) and more general risk parameterizations.(30,31) Time-to-event t can be any one of a number of time scales, including age, follow-up time, time since first exposure to the risk factor under study, or calendar year.”
BackgroundChoice of time scale and its effect on significance of predictors in longitudinal studies
Pencina, M. J., Larson, M. G., & D'Agostino, R. B. (2007) · Statistics in Medicine · doi:10.1002/sim.2699 · Cited by 183 · cited 1×
“age at entry (eg, a linear-quadratic model or stratification) reduced bias in estimated risk parameters.(14) However, with-”
Religion and assisted and non-assisted suicide in Switzerland: National Cohort Study
Spoerri, A., Zwahlen, M., Bopp, M., Gutzwiller, F., Egger, M., & Study, F. T. S. N. C. (2010) · International Journal of Epidemiology · doi:10.1093/ije/dyq141 · Cited by 137 · cited 2×
“”(16,19) Papers generally did not describe the numerical scale of the age adjustment variable (categorical or continuous) or what model was used, except that one paper mentioned including “age (years, continuous)” in a multivariate model,(18) and another mentioned using 10-year age categories.(15) None of the papers p…”
BackgroundMethodTime Scale and Adjusted Survival Curves for Marginal Structural Cox Models
Westreich, D. J., Cole, S. R., Tien, P. C., Chmiel, J. S., Kingsley, L., Funk, M. J., Anastos, K., & Jacobson, L. P. (2010) · American Journal of Epidemiology · doi:10.1093/aje/kwp418 · Cited by 73 · cited 1×
“related to risk was used as the primary time scale rather than time on study.(39) It may be somewhat extreme to say that age should always be used as the primary time scale in observational”
MethodHalanych, J. H., Safford, M. M., Kertesz, S. G., Pletcher, M. J., Kim, Y.-I., Person, S. D., Lewis, C. E., & Kiefe, C. I. (2010) · American Journal of Epidemiology · doi:10.1093/aje/kwp417 · Cited by 66
A prospective study of body mass index and mortality in Bangladesh
Pierce, B. L., Kalra, T., Argos, M., Parvez, F., Chen, Y., Islam, T., Ahmed, A., Hasan, R., Rakibuz-Zaman, M., Graziano, J., et al. (2010) · International Journal of Epidemiology · doi:10.1093/ije/dyp364 · Cited by 61 · cited 1×
“For example, one paper presented a plot of cumulative hazards “adjusted for age” and “plotted against time”; it can only be assumed that “time” was follow-up time and “age” was grouped age at entry.(21) Another paper mentioned using Cox proportional hazards models “stratified by age and time period for all analyses” a…”
MethodBressler, J., Folsom, A. R., Couper, D. J., Volcik, K. A., & Boerwinkle, E. (2010) · American Journal of Epidemiology · doi:10.1093/aje/kwp377 · Cited by 54
Melchior, M., Ferrie, J. E., Alexanderson, K. A. E., Goldberg, M., Kivimäki, M., Singh‐Manoux, A., Vahtera, J., Westerlund, H., Zins, M., & Head, J. A. (2010) · American Journal of Epidemiology · doi:10.1093/aje/kwq186 · Cited by 51
Age at diagnosis and the choice of survival analysis methods in cancer epidemiology
Cheung, Y. B., Gao, F., & Khoo, K. S. (2003) · Journal of Clinical Epidemiology · doi:10.1016/s0895-4356(02)00536-x · Cited by 49 · cited 1×
“risks being estimated and potential confounders to be adjusted.(12,25,38) Many applications of proportional hazards re-”
BackgroundAlcoholic Beverages and Prostate Cancer in a Prospective US Cohort Study
Watters, J. L., Park, Y., Hollenbeck, A. R., Schatzkin, A. G., & Albanes, D. (2010) · American Journal of Epidemiology · doi:10.1093/aje/kwq200 · Cited by 49
Analysis and Presentation of Data on the Natural Course of otitis media with effusion in Children
Zielhuis, G. A., Straatman, H., Rach, G. H., & van den Broek, P. (1990) · International Journal of Epidemiology · doi:10.1093/ije/19.4.1037 · Cited by 38
Liestøl, K., & Andersen, P. K. (2002) · Statistics in Medicine · doi:10.1002/sim.1322 · Cited by 33
Croezen, S., Haveman‐Nies, A., Picavet, H. S. J., Smid, E., de Groot, C. P. G. M., Veer, P. V., & Verschuren, W. M. M. (2010) · American Journal of Epidemiology · doi:10.1093/aje/kwq095 · Cited by 28
Gail, M. H., Graubard, B., Williamson, D. F., & Flegal, K. M. (2009) · Statistics in Medicine · doi:10.1002/sim.3473 · Cited by 27 · cited 1×
“With age as the primary time scale using left truncation on age at entry, it is generally not necessary to further adjust for age at entry unless entry into the study depends on factors related to the outcome.(32) Most statistical packages readily allow the use of age as the primary time scale, requiring only that the…”
MethodKabat, G. C., Kim, M., Kakani, C., Tindle, H., Wactawski‐Wende, J., Ockene, J. K., Luo, J., Wassertheil-Smoller, S. W., & Rohan, T. E. (2010) · American Journal of Epidemiology · doi:10.1093/aje/kwq159 · Cited by 27
A Prospective Study of Magnesium and Iron Intake and Pancreatic Cancer in Men
Kesavan, Y., Giovannucci, E. L., Fuchs, C. S., & Michaud, D. S. (2009) · American Journal of Epidemiology · doi:10.1093/aje/kwp373 · Cited by 26 · cited 1×
“Another paper mentioned using Cox proportional hazards models “stratified by age and time period for all analyses” and, in addition, “adjusted for potential confounders, including age” in multivariate analyses, which leaves us in the dark as to how the proportional hazards models were defined.(22) Most papers mentione…”
MethodRE: "TIME-TO-EVENT ANALYSIS OF LONGITUDINAL FOLLOW-UP OF A SURVEY: CHOICE OF THE TIME-SCALE"
Ingram, D. D., Makuc, D. M., & Feldman, J. J. (1997) · American Journal of Epidemiology · doi:10.1093/oxfordjournals.aje.a009309 · Cited by 12 · cited 1×
“Using follow-up time as the time scale in proportional hazards regression is not inherently incorrect.(37) The issue”
MethodEvaluating Bias From Birth-cohort Effects in the Age-based Cox Proportional Hazards Model
Hein, M. J., Schubauer‐Berigan, M. K., & Deddens, J. A. (2011) · Epidemiology · doi:10.1097/ede.0b013e3182093912 · Cited by 7
Cited by
Development and validation of the American Heart Association’s PREVENT equations
Khan, S. S., Matsushita, K., Sang, Y., Ballew, S. H., Grams, M. E., Surapaneni, A., Blaha, M. J., Carson, A. P., Chang, A. R., Ciemins, E., et al. (2024) · Circulation · doi:10.1161/circulationaha.123.067626 · Cited by 1,234
“57,58 Models were additionally adjusted for left truncation (people entering the study at different ages) as recommended in previous publications when age is used as the time scale.”
MethodKhan, S. S., Coresh, J., Pencina, M. J., Ndumele, C. E., Rangaswami, J., Chow, S. L., Palaniappan, L. P., Sperling, L. S., Virani, S. S., Ho, J. E., et al. (2023) · Circulation · doi:10.1161/cir.0000000000001191 · Cited by 797
Wadström, B. N., Wulff, A. B., Pedersen, K. M., Jensen, G. B., & Nordestgaard, B. G. (2022) · European Heart Journal · doi:10.1093/eurheartj/ehab705 · Cited by 234
Cytomegalovirus infection is associated with increased mortality in the older population
Savva, G. M., Pachnio, A., Kaul, B., Morgan, K., Huppert, F. A., Brayne, C., & Moss, P. A. H. (2013) · Aging Cell · doi:10.1111/acel.12059 · Cited by 216
“(Thi ebaut & B enichou, 2004; Cologne et al., 2012).”
BackgroundWadström, B. N., Pedersen, K. M., Wulff, A. B., & Nordestgaard, B. G. (2023) · European Heart Journal · doi:10.1093/eurheartj/ehac822 · Cited by 200
Metadata
- Authors
- Cologne, J., Hsu, W.-L., Abbott, R. D., Ohishi, W., Grant, E. J., Fujiwara, S., & Cullings, H. M.
- Affiliations
- Radiation Effects Research Foundation (JP)
- Center for Clinical Studies (US)
- Corresponding author
- Cologne, John
- Venue
- Epidemiology
- Published
- 2012-07
- Volume
- 23
- Issue
- 4
- Pages
- 565-573
- ISSN
- 1044-3983 (ISSN-L)
- OpenAlex
- W2074954887
- Cited by
- 127
- References
- 43
- Citations by year
- 2026: 5 · 2025: 4 · 2024: 10 · 2023: 17 · 2022: 12 · 2021: 16 · 2020: 7 · 2019: 5 · 2018: 10 · 2017: 4 · 2016: 13 · 2015: 11 · 2014: 2 · 2013: 7 · 2012: 3
- FWCI
- 7.03
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Epidemiology (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Inference
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Keywords
- confounding by age, Cox proportional hazards model, age-specific incidence, model adequacy, risk estimation bias
- MeSH terms
- Adult, Age Factors, Aged, Aged, 80 and over, Follow-Up Studies, Humans, Middle Aged, Neoplasms, Radiation-Induced, Time Factors, Bias, Confounding Factors, Epidemiologic, Multivariate Analysis, Proportional Hazards Models, Nuclear Weapons
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.