Epidemiology quiz

goergesandy2023

EvaluatingScreeningPrograms.pdf

Home >Applied Sciences homework help >Epidemiology quiz

Accounting for the Natural History of Disease in Screening

David Celentano, ScD, MHS Johns Hopkins University

Assessing the Effectiveness of Screening Programs Using Operational (Process) Measures

1.  Number of people screened

2.  Proportion of target population screened (number of times)

3.  Detected prevalence of preclinical disease

4.  Total costs of the program

5.  Costs per case found

6.  Costs per previously unknown case found

7.  Proportion of positive persons brought to final diagnosis and treatment

8.  Predictive value of a positive test in population screened

Source: adapted from Hulka BS. (1988). Cancer;62:1776–80. 2

Assessing the Effectiveness of Screening Programs Using Outcome Measures

1.  Reduction of mortality in the population screened

2.  Reduction of case-fatality in screened individuals

3.  Increase in percent of cases detected at earlier stages

4.  Reduction in complications

5.  Prevention of or reduction in recurrences

6.  Improvement of quality of life in screened individuals

Outcome

Biologic Onset of Disease

Therapy Diagnosis Symptoms

Natural History of Disease

An Example from HIV/AIDS

!  A critical point is a point in the natural history of the disease before which therapy may be less difficult and/or more effective

!  If the disease is potentially curable, cure may be possible before this point but not after

Critical Point

HIV Transmission Events

Natural History of Disease

!  Number of people screened positive

!  Number of people screened positive who were previously not known to be positive

!  Number of previously unknown positives who ultimately benefit

Assessing the Yield of a Screening Program

Possible Outcomes of a Screening Program

1.  All or most clinical cases of a disease first go through a detectable preclinical phase

2.  In the absence of intervention, all or most cases in a preclinical phase progress to a clinical phase

Assumptions Underlying a Relationship of Improved Outcome to Early Detection of Disease

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

Epidemiologic Designs to Address Early Detection of Disease

Section B

Natural History of Cervical Cancer

Normal In situ Carcinoma

Invasive cancer

Design of a Non-Randomized Comparison (Cohort) Study of the Benefits of Screening

Die from the disease

Don’t die from the disease

Die from the disease

Don’t die from the disease

Screened Not screened

1.  Selection bias a.   Referral (volunteer) bias

Problems Which Complicate Assessment of Improvement in Survival as a Result of Early Detection

1.  Selection bias a.  Referral (volunteer) bias b.   Length-biased sampling

Problems Which Complicate Assessment of Improvement in Survival as a Result of Early Detection

Natural History

1.  Selection bias a.  Referral (volunteer) bias b.  Length-biased sampling c.   Lead-time bias

Problems Which Complicate Assessment of Improvement in Survival as a Result of Early Detection

!  Lead time: interval by which the time of diagnosis can be advanced by the screening procedure, as compared with the normal methods for detection and diagnosis

Lead Time

Five-Year Survival and Lead-Time Bias

Lead Time Bias: Five-Year Survival When Diagnosis Is Made Without Screening

Lead Time Bias: Shift of Five-Year Period by Screening and Early Detection (Lead Time)

Lead Time Bias: Bias in Survival Calculation Resulting from Early Detection

Design of a Randomized Trial of the Benefits of Screening: HIP Study

HIP enrollees ~ 62,000

Screening including mammography

~31,000

Compare mortality

Regular care ~31,000

R a n d o m i z e d

Breast cancer

No Breast cancer

Breast cancer

No Breast cancer

Deaths Due to Breast Cancer HIP Five-Year Follow-Up

Case Fatality Due to Breast Cancer HIP Five-Year Follow-Up

!  The apparent lack of benefit may be inherent in the natural history of the disease

!  The therapeutic intervention may not be any more effective when it is provided earlier in the natural history of disease

!  Inadequacies in the care provided may account for the observed lack of benefit

Interpreting Results that Show No Benefit of Screening

Cervical Cancer Example

Section C

Cancer Death Rates* among US Women, 1930– 2005

Source: adapted from: http://www.cancer.org/; US Mortality Data 1960–2005, US Mortality Volumes 1930–1959, National Center for Health Statistics, Centers for Disease Control and Prevention, 2008. 2

Cervical cancer and cancer of the body of the uterus

Cancer Death Rates* among US Women, 1930– 2005

Cervical cancer and cancer of the body of the uterus

Cancer Death Rates* among US Women, 1930– 2005

Burden time

Cervical Cancer Is Not a Top-10 Cause, 2010

!  Cervical cancer deaths: 4,210

!  2010 estimated deaths* !  Lung and bronchus: 26% !  Breast: 15% !  Colon and rectum: 9% !  Pancreas: 7% !  Ovary: 5% !  Non-Hodgkin lymphoma: 4% !  Leukemia: 3% !  Uterine corpus: 3% !  Liver and intrahepatic bile duct: 2% !  Brain/ONS: 2% !  All other sites: 24%

ONS=Other nervous system. *Excludes basal and squamous cell skin cancers and in situ carcinomas except urinary bladder Source: adapted from: http://www.cancer.org/ 7

Uterine corpus: 3%

4,210

Cervical Cancer Is Not a Top-10 Cause, 2010

!  Cervical cancer deaths: 4,210

!  Cervical cancer cases: 12,200

!  2010 estimated deaths* !  Breast: 28% !  Lung and bronchus: 14% !  Colon and rectum: 10% !  Uterine corpus: 6% !  Thyroid: 5% !  Non-Hodgkin lymphoma: 4% !  Melanoma of skin: 4% !  Kidney and renal pelvis: 3% !  Ovary: 3% !  Pancreas: 3% !  All other sites: 20%

ONS=Other nervous system. *Excludes basal and squamous cell skin cancers and in situ carcinomas except urinary bladder Source: adapted from: http://www.cancer.org/ 8

Uterine corpus: 6%

12,200

Leading Sites of New Cancer Cases and Deaths Worldwide by Level of Economic Development, 2007—Developed Countries, Females

Source: adapted from: Global Cancer Facts & Figures, 2007, http://www.cancer.org/ 9

Cervix uteri 42,101

Cervix uteri 87,466

Estimated deaths Estimated new cases

Leading Sites of New Cancer Cases and Deaths Worldwide by Level of Economic Development, 2007—Worldwide, Females

Source: adapted from: Global Cancer Facts & Figures, 2007, http://www.cancer.org/ 10

Cervix uteri 309,808

Cervix uteri 555,094

Estimated deaths Estimated new cases

#2 cancer in women worldwide

Leading Sites of New Cancer Cases and Deaths Worldwide by Level of Economic Development, 2007—Developing Countries, Females

Source: adapted from: Global Cancer Facts & Figures, 2007, http://www.cancer.org/ 11

Cervix uteri 272,238

Cervix uteri 473,430

Estimated deaths Estimated new cases

International Variation in Age-Standardized Cervical Cancer Incidence Rates

Sources: Global Cancer Facts & Figures, 2007 http://www.cancer.org/; GLOBOCAN 2002, http://www-dep.iarc.fr/ 12

!  Make observations based on: !  Variation in incidence and mortality rates by person, place, and time !  Case reports, case series !  Correlation of population-level exposures with population incidence or mortality

rates !  Simple comparisons of groups of individuals with differing characteristics

!  Coupled with existing biomedical knowledge

Generating Hypotheses about the Causes of the Public Health Problem

!  1842: Rigoni-Stern (Italy) !  Reported that the ratio of deaths from breast cancer to deaths from cancer of the

womb was far greater in nuns (and single women) than in other women

Simple Observations

14 Source: Griffiths M. (1991). Nuns, virgins, and spinsters. Rigoni-Stern and cervical cancer revisited. Br J Obstet Gynaecol;98:797–802.

!  1901: Braithwaite (England) !  Reported in the Lancet that cancer of the cervix:

•  “Was seldom or never met with amongst the numerous Jewesses” •  The difference in diet between Jews and gentiles consists mainly in the absence

of bacon and ham from the diet of Jews” (salt or pork hypothosis)

Simple Observations

15 Source: Griffiths M. (1991). Nuns, virgins, and spinsters. Rigoni-Stern and cervical cancer revisited. Br J Obstet Gynaecol;98:797–802.

“A case of carcinoma of the cervix in a young woman following gonorrhea The following notes refer to a young II-para, aged 28, in whom carcinoma of the cervix followed an attack of gonorrhea contracted five months previously. The occurrence of a carcinomatous growth during local treatment of the cervical canal for gonococcal cervicitis seems to support the ‘irritation’ theory in that in this case there was the discharge and also the repeated application of antiseptics.”

—S. Gordon Luker, MD

Lancet, July 8, 1922

Making Observations: Case Report, 1922

!  1949: Versluys, the Netherlands !  Reports that deaths from cervical cancer were more

common in married women than unmarried (5.92% vs. 2.33%)

!  Cervical cancer is no less common in nuns than in other single women

!  1969: Fraumeni, United States !  Reported that the lower cervical cancer rates for nuns

than for the US population of women seemed related to coital factors (Fraumeni JF et al., 1969, J Natl Cancer Inst;42:455–68)

Making Observations: Simple and not So

17 Source: Griffiths M. (1991). Nuns, virgins, and spinsters. Rigoni-Stern and cervical cancer revisited. Br J Obstet Gynaecol;98:797–802.

!  Be specific about: !  The population at risk (who, where, and when) !  The exposure !  The outcome

Testing Hypotheses about the Causes: State the Research Question for your Study

!  “Do women aged 20–35 years living in Washington state between 1960 and 1970 who have had a greater number of sex partners subsequently have a higher risk of cervical cancer?”

!  The population at risk !  Who: women aged 20–35 years !  Where: living in Washington state !  When: between 1960 and 1970

!  The exposure: those who have had a greater number of sex partners

!  The outcome: cervical cancer

For Example: “Coital Factors”

!  Epidemiology encompasses design and analytical methods used to investigate disease occurrence in human populations

!  These methods can be used to address a wide range of hypotheses in biomedical research

!  Testing a hypothesis generally involves comparing the occurrence of the endpoint in two or more groups that differ on a factor of interest (may lead to a causal inference)

Testing Hypotheses Using Epidemiology

The End of the Story

Section D

Case-Control Study in Jewish Women

Source: Martin CE. (1967). Am J Public Health; 57:803–814.

Cases Controls

!  Martin inferred that the same mechanism of cervical cancer causation was operating from population to population !  Differences in risk more likely due to differences in the prevalence of the causal

factor

Inferring a Causal Model for Cervical cancer

“… It is proposed that squamous-cell cervical carcinoma shares many characteristics in common with communicable diseases, which follow a venereal mode of transmission.”

Martin’s “Interpretation and Conclusions”

!  Ask husbands about their own sexual history in a study of women who had no sexual partner other than their husband

Does Sex Cause Cervical Cancer?

5 Source: Buckley JD et al. (1981). Lancet;2:1010–1015.

Husband Overall

RR Cervical cancer

Carcinoma in situ/dysplasia

Number of sex partners

1 1.00 1.00 1.00

2–5 1.65 1.20 2.13

6–15 2.63 2.42 2.74

16+ 7.82 6.76 9.84

P-trend <0.01 <0.05 <0.05

�We are unable to think of any interpretation of these data other than the obvious one that men can contract and transmit to their wives a venereal infection which increases the risk of premalignant and malignant change in their cervical epithelium.”

Drawing Inferences

Source: Buckley JD et al. (1981). Lancet;2:1010–1015. 6

!  Sex does not cause cervical cancer directly

Drawing Inferences

7 Source: Buckley JD et al. (1981). Lancet;2:1010–1015.

Sex Cervical cancer

Transmissible agent ?

!  Herpes simplex virus 2 (HSV-2)?

What Is the Infectious Causal Factor?

8 Source: S Sprecher-Goldberger et al. (1970). Lancet; 2:266.

Cases: cervical Cancer

Controls: other GYN diseases

Controls: other cancers

HSV-2 antibody positive

83% 29% 33%

“The similarity of the last two figures indicates clearly that patients from different hospitals had been exposed to the same risk of genital herpetic infection, and that the high incidence of this infection is genuinely localised among patients with carcinoma of the cervix.”

Authors’ Conclusions

Source: S Sprecher-Goldberger et al. (1970). Lancet; 2:266. 9

Cases: cervical Cancer

Controls: other GYN diseases

Controls: other cancers

HSV-2 antibody positive

83% 29% 33%

!  Consider the biological knowledge available at the time !  Was there any evidence that HSV-2 is oncogenic? no !  Were there other viral, bacterial, or other types of infectious agents that are known

to be carcinogenic? yes

What Is the Infectious Causal Factor?

!  1974: zur Hausen !  Hypothesized the role of human papillomavirus (HPV) in cervical cancer after rabbit

studies

!  1983: Durst, Gissmann, Ikenberg, and zur Hausen !  Isolated HPV 16 DNA from an invasive cervical cancer

!  1992: Muñoz et al. !  In a large epidemiologic study, found that high-risk HPV types are the primary risk

factor for cervical cancer

!  1995: International Agency for Research on Cancer (IARC) !  HPV 16/18 determined to cause cervical cancer

Making Observations: Biology and Epidemiology

11 Source: H Zur Hausen. (2002). Lancet;2:342–349.

!  1995: Bosch et al. !  Showed that 93% of

cervical cancers contained HPV DNA in a large international consortium study

!  1999: Walboomers et al. !  With improved DNA

detection techniques, showed that 99.7% of cervical cancers contained HPV DNA

Making Observations: Biology and Epidemiology

Nobel Prize in Physiology or Medicine, 2008

!  Harald zur Hausen

!  “The Search for Infectious Causes of Human Cancers: Where and Why?”

!  See lecture page for link to “Banquet Speech by Harald zur Hausen”

Establishing Association: Key Epi Studies

Sex HPV infection

Pre-malignant lesions of the cervix

HPV infection

Does HPV Cause Cervical Pre-Neoplastic Lesions?

Source: Schiffman MH et al. (1993). J Natl Cancer Inst;85:958–964. 15

HPV test result RR* 95% CI

Negative 1.0 Ref

Types 6, 11, 42, other, unknown 8.7 5.8–13

Types 31, 33, 35, 39, 45, 51, 52 33 18–59

Types 16 or 18 51 28–94

* Relative risk; results were the same after adjusting for known risk factors for cervical cancer

!  For a factor to cause an outcome, it necessarily must precede the existence of the outcome

Temporality Is Necessary for Causal Claims

A B

Sex Cervical cancer

Transmissible agent

!  Does the exposure precede the outcome?

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Does the exposure precede the outcome? yes

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Does the exposure precede the outcome? yes

!  Is there a plausible mechanism that may explain the observed association between the exposure and the outcome?

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Does the exposure precede the outcome? yes

!  Is there a plausible mechanism that may explain the observed association between the exposure and the outcome? yes

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Does the exposure precede the outcome? yes

!  Is there a plausible mechanism that may explain the observed association between the exposure and the outcome? yes

!  How strong is the association?

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Does the exposure precede the outcome? yes

!  Is there a plausible mechanism that may explain the observed association between the exposure and the outcome? yes

!  How strong is the association? very strong

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Does the exposure precede the outcome? yes

!  Is there a plausible mechanism that may explain the observed association between the exposure and the outcome? yes

!  How strong is the association? very strong

!  How consistent is the association among studies evaluating the exposure-outcome association?

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Does the exposure precede the outcome? yes

!  Is there a plausible mechanism that may explain the observed association between the exposure and the outcome? yes

!  How strong is the association? very strong

!  How consistent is the association among studies evaluating the exposure-outcome association? very consistent

Drawing Inferences about Exposure-Outcome Associations

HPV infection

Cervical cancer

!  Listed some criteria to evaluate screening effectiveness

!  Described the natural history of disease and how it influences screening

!  Discussed alternative designs to evaluate screening program and outcomes

!  Evaluated sources of bias in screening

!  Discussed the history of screening for cervical cancer

Summary

Your feedback is very important and will be used for future revisions. The Evaluation link is available on the lecture page.

Lecture Evaluation