Get questions and answers for Fundamentals of Biostatistics

GET Fundamentals of Biostatistics TEXTBOOK SOLUTIONS

1 Million+ Step-by-step solutions
math books
1. Use logistic regression methods to assess whether there is an association between ovarian cancer risk and duration of OC use while controlling for age. Provide a two-sided p-value. Assume that the average duration of use in the < 3 years group = 1.5 years and in the 3+ years group = 4 years. Also, provide an estimate of the OR relating ovarian cancer risk per year of use of OCs and a 95% CI?

2. Use logistic regression methods to assess whether there is an association between ever use of OCs and ovarian cancer risk, while controlling for age. Also, provide an estimate of the OR and a 95% CI about this estimate. 


Cancer

A case–control study was performed early in the Nurses’ Health Study (NHS) to assess the possible association between oral contraceptive (OC) use and ovarian cancer [50]. Forty seven ovarian cancer cases were identified at or before baseline (1976). For each case, 10 controls matched by year of birth and with intact ovaries at the time of the index woman’s diagnosis were randomly chosen from questionnaire respondents free from ovarian cancer. The data in Table 13.58 were presented.

Table 13.58 Duration of OC use by age at diagnosis among women with ovarian cancer and controls

 

What is the attributable risk (AR) of high BMI (≥25) for progression of AMD? 


Ophthalmology

The data in Table 13.59 were presented relating body mass index (BMI) to progression of advanced age-related macular degeneration (AMD), a common eye disease in the elderly that results in significant visual loss [51].

Table 13.59 Association between BMI and progression of AMD

 

Provide a 95% CI about this estimate?


Ophthalmology

The data in Table 13.59 were presented relating body mass index (BMI) to progression of advanced age-related macular degeneration (AMD), a common eye disease in the elderly that results in significant visual loss [51].

Table 13.59 Association between BMI and progression of AMD

 

1. Obtain the crude OR estimate, and provide a 95% CI for the crude OR?

2. Test the null hypothesis of no association between aspirin assignment and CVD?


Cardiovascular Disease

The Women’s Health Study randomly assigned 39,876 initially healthy women ages 45 years or older to receive either 100 mg of aspirin on alternate days or placebo and monitored them for 10 years for a major cardiovascular event [52]. Table 13.60 shows the results stratified by age at randomization.

Table 13.60 Incidence of CVD by treatment group and age in the Women’s Health Study

 


Use logistic regression methods to characterize the relationship between aspirin assignment and the odds of CVD, by doing the following.

Evaluate whether age confounds the CVD− aspirin relationship by using dummy variables for age categories; calculate the age-adjusted OR estimate and 95% CI?


Cardiovascular Disease

The Women’s Health Study randomly assigned 39,876 initially healthy women ages 45 years or older to receive either 100 mg of aspirin on alternate days or placebo and monitored them for 10 years for a major cardiovascular event [52]. Table 13.60 shows the results stratified by age at randomization.

Table 13.60 Incidence of CVD by treatment group and age in the Women’s Health Study

 


Use logistic regression methods to characterize the relationship between aspirin assignment and the odds of CVD, by doing the following.

Evaluate whether age is an effect modifier of the relationship between aspirin and CVD?


Cardiovascular Disease

The Women’s Health Study randomly assigned 39,876 initially healthy women ages 45 years or older to receive either 100 mg of aspirin on alternate days or placebo and monitored them for 10 years for a major cardiovascular event [52]. Table 13.60 shows the results stratified by age at randomization.

Table 13.60 Incidence of CVD by treatment group and age in the Women’s Health Study

 


Use logistic regression methods to characterize the relationship between aspirin assignment and the odds of CVD, by doing the following.

Treat testosterone as a continuous variable (suitably transformed if necessary)?

Repeat the analyses in Problems 13.122–13.124 using diastolic bp (DBP) z score group defined in the same manner as SBP z-score group?


Hypertension

The data set WALES.DAT contains familial data on blood pressure (bp) in two-communities in South Wales (the Rhondda Fach and the Vale of Glamorgan). Subjects were seen at 4 visits from the mid 1950s to the early 1960s. For this problem we will focus on bp among adults (age ≥ 30) at the first visit. To standardize bp for age and sex, we will use the z-score approach within 10-year age-sex groups (i.e., age 30–39 males/40–49 males/50–59 males/60+ males/30–39 females/40–49 females/50–59 females/60+ females) where

 


We then group SBP z-score for ease of interpretation as

follows:

SBP z-score group = 2 if SBP z-score ≥ 1.0,

= 1 if SBP z-score ≥ 0.5 and < 1.0,

= 0 if SBP z-score < 0.5 and not missing

Compare your results in Problems 13.122 and 13.123?


Hypertension

The data set WALES.DAT contains familial data on blood pressure (bp) in two-communities in South Wales (the Rhondda Fach and the Vale of Glamorgan). Subjects were seen at 4 visits from the mid 1950s to the early 1960s. For this problem we will focus on bp among adults (age ≥ 30) at the first visit. To standardize bp for age and sex, we will use the z-score approach within 10-year age-sex groups (i.e., age 30–39 males/40–49 males/50–59 males/60+ males/30–39 females/40–49 females/50–59 females/60+ females) where

 


We then group SBP z-score for ease of interpretation as

follows:

SBP z-score group = 2 if SBP z-score ≥ 1.0,

= 1 if SBP z-score ≥ 0.5 and < 1.0,

= 0 if SBP z-score < 0.5 and not missing

Use ordinal logistic regression to assess the association between SBP z-score group and the same variables in Problem 13.122?


Hypertension

The data set WALES.DAT contains familial data on blood pressure (bp) in two-communities in South Wales (the Rhondda Fach and the Vale of Glamorgan). Subjects were seen at 4 visits from the mid 1950s to the early 1960s. For this problem we will focus on bp among adults (age ≥ 30) at the first visit. To standardize bp for age and sex, we will use the z-score approach within 10-year age-sex groups (i.e., age 30–39 males/40–49 males/50–59 males/60+ males/30–39 females/40–49 females/50–59 females/60+ females) where

 


We then group SBP z-score for ease of interpretation as

follows:

SBP z-score group = 2 if SBP z-score ≥ 1.0,

= 1 if SBP z-score ≥ 0.5 and < 1.0,

= 0 if SBP z-score < 0.5 and not missing

13.122 Use polytomous logistic regression with group = 0 as the reference group to assess the association between SBP z-score group and the following variables:

(a) BMI,

(b) height, 

(c) region (i.e., Rhondda Fach vs. Vale of Glamorgan),

(d) occupation (treat occupation codes of 5–9 as missing for this analysis). Of particular interest is the comparison of group = 2 vs. group = 0. 


Hypertension

The data set WALES.DAT contains familial data on blood pressure (bp) in two-communities in South Wales (the Rhondda Fach and the Vale of Glamorgan). Subjects were seen at 4 visits from the mid 1950s to the early 1960s. For this problem we will focus on bp among adults (age ≥ 30) at the first visit. To standardize bp for age and sex, we will use the z-score approach within 10-year age-sex groups (i.e., age 30–39 males/40–49 males/50–59 males/60+ males/30–39 females/40–49 females/50–59 females/60+ females) where

 


We then group SBP z-score for ease of interpretation as

follows:

SBP z-score group = 2 if SBP z-score ≥ 1.0,

= 1 if SBP z-score ≥ 0.5 and < 1.0,

= 0 if SBP z-score < 0.5 and not missing

Assess whether the effect of age at menarche is different for Caucasian vs. African American women. Report a p-value (two-tailed)?


Two logistic models were run with Stata Version 11 using these data. For the 1st model (Table 13.62) we fit

Logit (Pi) = Î± + Î²1x1 + Î²2x2

Where x1 = age at menarche (1 represents ≥ 12, 0 represents < 12),

X2 = race (1 = African American, 0 Caucasian).

For the 2nd model (Table 13.63) we fit

Logit (Pi) = Î± + Î²1x1 + Î²2x2 + Î²3x1x2

Table 13.62 Logistic regression of ovarian cancer on age at menarche and race

 


Table 13.63 Logistic regression of ovarian cancer on age at menarche, race, and age at menarche × race

What does the variable β1mean in the 2nd logistic regression (Table 13.63)? How does it differ from the meaning of β1in the 1st logistic regression (Table 13.62)? 


Two logistic models were run with Stata Version 11 using these data. For the 1st model (Table 13.62) we fit

Logit (Pi) = Î± + Î²1x1 + β2x2

Where x1 = age at menarche (1 represents ≥ 12, 0 represents < 12),

X2 = race (1 = African American, 0 Caucasian).

For the 2nd model (Table 13.63) we fit

Logit (Pi) = Î± + Î²1x1 + β2x2 + β3x1x2

Table 13.62 Logistic regression of ovarian cancer on age at menarche and race

 


Table 13.63 Logistic regression of ovarian cancer on age at menarche, race, and age at menarche × race

 

Estimate the OR for the association between age at menarche and ovarian cancer after controlling for race, and provide a 95% CI about this estimate?


Two logistic models were run with Stata Version 11 using these data. For the 1st model (Table 13.62) we fit

Logit (Pi) = Î± + Î²1x1 + β2x2

Where x1 = age at menarche (1 represents ≥ 12, 0 represents < 12),

X2 = race (1 = African American, 0 Caucasian).

For the 2nd model (Table 13.63) we fit

Logit (Pi) = Î± + Î²1x1 + β2x2 + β3x1x2

Table 13.62 Logistic regression of ovarian cancer on age at menarche and race

 


Table 13.63 Logistic regression of ovarian cancer on age at menarche, race, and age at menarche × race

 

For both Caucasians and African Americans (seperately), estimate the OR between late age at menarche (≥12) and ovarian cancer risk and provide a 95% CI about these estimate?


Cancer

Results from a population-based case–control study of ovarian cancer were recently reported from the North Carolina Case–Control Study based on data collected from 1999–2008 [53]. Cases were women with ovarian cancer who were ages 20–74 from 48 North Carolina counties; controls were frequency matched by age and race and were recruited from the same geographic regions using randomdigit dialing. Controls could not have a bilateral oophorectomy. The data in Table 13.61 were reported concerning the association between age at menarche (age when periods begin) and ovarian cancer.

Table 13.61 Association between age at menarche and ovarian cancer

 

Discuss your results from Problems 13.115 and 13.116?

Treat testosterone as a categorical variable in quartiles, with the 1st quartile as the reference group?

Compare the incidence density of breast cancer in current users vs. never users, and report a p-value?


Cancer

The data relating oral-contraceptive (OC) use and the incidence of breast cancer in the age group 40−44 in the NHS are given in Table 14.31.

Table 14.31: Relationship between breast-cancer incidence and OC use among 40- to 44-year-old women in the NHS

 

Compare the incidence density of breast cancer in past users vs. never users, and report a p-value?


Cancer

The data relating oral-contraceptive (OC) use and the incidence of breast cancer in the age group 40−44 in the NHS are given in Table 14.31.

Table 14.31: Relationship between breast-cancer incidence and OC use among 40- to 44-year-old women in the NHS

 

Estimate the rate ratio comparing current users vs. never users, and provide a 95% CI about this estimate?


Cancer

The data relating oral-contraceptive (OC) use and the incidence of breast cancer in the age group 40−44 in the NHS are given in Table 14.31.

Table 14.31: Relationship between breast-cancer incidence and OC use among 40- to 44-year-old women in the NHS

 

Estimate the rate ratio comparing past users vs. never users, and provide a 95% CI about this estimate?


Cancer

The data relating oral-contraceptive (OC) use and the incidence of breast cancer in the age group 40−44 in the NHS are given in Table 14.31.

Table 14.31: Relationship between breast-cancer incidence and OC use among 40- to 44-year-old women in the NHS

 

How much power did the study have for detecting an IRR for breast cancer of 1.5, comparing current OC users vs. never OC users among 40- to 44-year-old women if

(a) The true incidence rate of breast cancer among never users and the amount of person-time for current and never users are the same as in Table 14.31,

(b) The expected number of events for never OC users is the same as the observed number of events in Table 14.31, and

(c) The average follow-up time per subject is the same for both current and never OC users? 


Cancer

The data relating oral-contraceptive (OC) use and the incidence of breast cancer in the age group 40−44 in the NHS are given in Table 14.31.

Table 14.31: Relationship between breast-cancer incidence and OC use among 40- to 44-year-old women in the NHS

 

What is the expected number of events that need to be realized in each group to achieve 80% power to detect an IRR for breast cancer of 1.5 for current OC users vs. never OC users under the same assumptions as in Problem 14.5?


Cancer

The data relating oral-contraceptive (OC) use and the incidence of breast cancer in the age group 40−44 in the NHS are given in Table 14.31.

Table 14.31: Relationship between breast-cancer incidence and OC use among 40- to 44-year-old women in the NHS

 

Compare survival curves of the two groups, using hypothesis-testing methods, and report a p-value?

Assess the significance of each of the variables?


A Cox proportional-hazards model was fit to these data to assess the relationship between age, sex, number of cigarettes smoked, and log10CO concentration, when considered simultaneously, on the ability to remain abstinent from smoking. The results are given in Table 14.32.

Table 14.32 Proportional-hazards model relating the hazard of recidivism to age, sex, number of cigarettes smoked prior to quitting, and log10CO concentration

 


aThis variable represents CO values adjusted for minutes elapsed since last cigarette smoked prior to quitting.

Estimate the effects of each variable in terms of hazard ratios, and provide 95% confidence limits corresponding to each point estimate?


A Cox proportional-hazards model was fit to these data to assess the relationship between age, sex, number of cigarettes smoked, and log10CO concentration, when considered simultaneously, on the ability to remain abstinent from smoking. The results are given in Table 14.32.

Table 14.32 Proportional-hazards model relating the hazard of recidivism to age, sex, number of cigarettes smoked prior to quitting, and log10CO concentration

 


aThis variable represents CO values adjusted for minutes elapsed since last cigarette smoked prior to quitting.

Compare the crude and adjusted analyses of the relationship of log10CO to recidivism in Problems 14.8 and 14.9?


A Cox proportional-hazards model was fit to these data to assess the relationship between age, sex, number of cigarettes smoked, and log10CO concentration, when considered simultaneously, on the ability to remain abstinent from smoking. The results are given in Table 14.32.

Table 14.32 Proportional-hazards model relating the hazard of recidivism to age, sex, number of cigarettes smoked prior to quitting, and log10CO concentration

 


aThis variable represents CO values adjusted for minutes elapsed since last cigarette smoked prior to quitting.

Suppose we regard a preparation as being bioavailable for a subject at the first week when level of plasma carotene increases by 50% from the baseline level (based on an average of the first and second baseline determinations). Use survival-analysis methods to estimate the proportion of subjects for whom the preparation is not bioavailable at different points in time?

Assess whether there are significant differences among the survival curves obtained in Problem 14.12?

Answer the same question as in Problem 14.12 assuming the criterion for bioavailability is a 100% increase in plasma-carotene level from baseline?

Answer the same question posed in Problem 14.13 assuming the criterion for bioavailability is a 100% increase in plasma-carotene level from baseline?

Estimate the hazard function by year for each group?


Ophthalmology

In Table 14.33, we present data from the RP clinical trial described in Example 14.30 (on page 813) concerning effect of high-dose vitamin E (400 IU/day) vs. low-dose vitamin E (3 IU/day) on survival (where failure is loss of at least 50% of initial ERG 30 Hz amplitude).

Table 14.33 Number of patients who failed, were censored, or survived by year in the 400 IU vitamin E group and 3 IU vitamin E group, respectively, RP clinical trial

 

Estimate the survival probability by year for each group?


Ophthalmology

In Table 14.33, we present data from the RP clinical trial described in Example 14.30 (on page 813) concerning effect of high-dose vitamin E (400 IU/day) vs. low-dose vitamin E (3 IU/day) on survival (where failure is loss of at least 50% of initial ERG 30 Hz amplitude).

Table 14.33 Number of patients who failed, were censored, or survived by year in the 400 IU vitamin E group and 3 IU vitamin E group, respectively, RP clinical trial

 

Obtain a 95% CI for the survival probability at year 6 for each group?


Ophthalmology

In Table 14.33, we present data from the RP clinical trial described in Example 14.30 (on page 813) concerning effect of high-dose vitamin E (400 IU/day) vs. low-dose vitamin E (3 IU/day) on survival (where failure is loss of at least 50% of initial ERG 30 Hz amplitude).

Table 14.33 Number of patients who failed, were censored, or survived by year in the 400 IU vitamin E group and 3 IU vitamin E group, respectively, RP clinical trial

 

Compare overall survival curves of the two groups, and obtain a p-value?


Ophthalmology

In Table 14.33, we present data from the RP clinical trial described in Example 14.30 (on page 813) concerning effect of high-dose vitamin E (400 IU/day) vs. low-dose vitamin E (3 IU/day) on survival (where failure is loss of at least 50% of initial ERG 30 Hz amplitude).

Table 14.33 Number of patients who failed, were censored, or survived by year in the 400 IU vitamin E group and 3 IU vitamin E group, respectively, RP clinical trial

 

Suppose a new study is planned, with 200 patients randomly assigned to each of a 400 IU per day vitamin E group and a 3 IU per day vitamin E group. If the survival experience in the 3 IU per day group is assumed the same as in Table 14.33, the relative hazard for the 400 IU/day group vs. the 3 IU/day group = 1.5, and the censoring experience of both groups are assumed the same as for the 3 IU per day group in Table 14.33, then how much power would a new study have if the maximum duration of follow-up is 4 years (rather than 6 years as in the original study) and a two-sided test is used with α = .05? 


Ophthalmology

In Table 14.33, we present data from the RP clinical trial described in Example 14.30 (on page 813) concerning effect of high-dose vitamin E (400 IU/day) vs. low-dose vitamin E (3 IU/day) on survival (where failure is loss of at least 50% of initial ERG 30 Hz amplitude).

Table 14.33 Number of patients who failed, were censored, or survived by year in the 400 IU vitamin E group and 3 IU vitamin E group, respectively, RP clinical trial

 

How many subjects need to be enrolled in each group (assume equal sample size in each group) to achieve 80% power if a two-sided test is used with α = .05 and the same assumptions are made as in Problem 14.20? 


Ophthalmology

In Table 14.33, we present data from the RP clinical trial described in Example 14.30 (on page 813) concerning effect of high-dose vitamin E (400 IU/day) vs. low-dose vitamin E (3 IU/day) on survival (where failure is loss of at least 50% of initial ERG 30 Hz amplitude).

Table 14.33 Number of patients who failed, were censored, or survived by year in the 400 IU vitamin E group and 3 IU vitamin E group, respectively, RP clinical trial

 

Implement the test in Problem 14.55, and report a p-value (two-tailed)?


Cancer

A study was performed to compare breast cancer incidence between postmenopausal women who used PMH vs. women who did not. A group of 200 women who were current PMH users and 1000 women who were never PMH users in 1990 in the NHS were identified. All women were postmenopausal and free of cancer as of 1990. The 1200 women were ascertained for incident breast cancer by mail questionnaire every 2 years up to the year 2000. However, not all women had complete follow-up. For simplicity, we will assume that women can only fail every 2 years, i.e., in 1992, 1994, . . . , 2000. The results are given in Table 14.36.

Table 14.36 Relationship between PMH use and breast cancer incidence among 1200 women in the NHS

 



aFailed means developed breast cancer.

bAssume that at any given year that the failures occur just prior to the censored observations in that year.

What test can be used to compare the incidence of breast cancer between the 2 groups, taking into account the time when breast cancer develops and the length of follow-up of each subject? 


Cancer

A study was performed to compare breast cancer incidence between postmenopausal women who used PMH vs. women who did not. A group of 200 women who were current PMH users and 1000 women who were never PMH users in 1990 in the NHS were identified. All women were postmenopausal and free of cancer as of 1990. The 1200 women were ascertained for incident breast cancer by mail questionnaire every 2 years up to the year 2000. However, not all women had complete follow-up. For simplicity, we will assume that women can only fail every 2 years, i.e., in 1992, 1994, . . . , 2000. The results are given in Table 14.36.

Table 14.36 Relationship between PMH use and breast cancer incidence among 1200 women in the NHS

 



aFailed means developed breast cancer.

bAssume that at any given year that the failures occur just prior to the censored observations in that year.

Estimate the 10-year incidence of breast cancer in each group?


Cancer

A study was performed to compare breast cancer incidence between postmenopausal women who used PMH vs. women who did not. A group of 200 women who were current PMH users and 1000 women who were never PMH users in 1990 in the NHS were identified. All women were postmenopausal and free of cancer as of 1990. The 1200 women were ascertained for incident breast cancer by mail questionnaire every 2 years up to the year 2000. However, not all women had complete follow-up. For simplicity, we will assume that women can only fail every 2 years, i.e., in 1992, 1994, . . . , 2000. The results are given in Table 14.36.

Table 14.36 Relationship between PMH use and breast cancer incidence among 1200 women in the NHS

 


aFailed means developed breast cancer.

bAssume that at any given year that the failures occur just prior to the censored observations in that year.

What does a censored observation in 1992 mean in the context of these data? 


Cancer

A study was performed to compare breast cancer incidence between postmenopausal women who used PMH vs. women who did not. A group of 200 women who were current PMH users and 1000 women who were never PMH users in 1990 in the NHS were identified. All women were postmenopausal and free of cancer as of 1990. The 1200 women were ascertained for incident breast cancer by mail questionnaire every 2 years up to the year 2000. However, not all women had complete follow-up. For simplicity, we will assume that women can only fail every 2 years, i.e., in 1992, 1994, . . . , 2000. The results are given in Table 14.36.

Table 14.36 Relationship between PMH use and breast cancer incidence among 1200 women in the NHS

 



aFailed means developed breast cancer.

bAssume that at any given year that the failures occur just prior to the censored observations in that year.

The age-adjusted rate ratio between the groups in Problem 14.51 was 0.34. Is this different from the estimated rate ratio in Problem 14.51? If so, why?


Health Promotion

A recent, article by Kenfield et al. [12] studied the relationship between various aspects of smoking and mortality among 104,519 women in the Nurses’ Health Study (NHS) from 1980−2004. One issue is whether there is a mortality benefit from quitting smoking vs. continuing to smoke and, if so, how long it takes for the mortality experience of former smokers to approximate that of never smokers. The data in Table 14.35 were presented comparing former smokers with current smokers.

Table 14.35: Relationship of time since quitting to total mortality

 

What is the estimated rate ratio for total mortality between former smokers who quit 20+ years ago and current smokers? Provide a 95% CI for this estimate?


Health Promotion

A recent, article by Kenfield et al. [12] studied the relationship between various aspects of smoking and mortality among 104,519 women in the Nurses’ Health Study (NHS) from 1980−2004. One issue is whether there is a mortality benefit from quitting smoking vs. continuing to smoke and, if so, how long it takes for the mortality experience of former smokers to approximate that of never smokers. The data in Table 14.35 were presented comparing former smokers with current smokers.

Table 14.35: Relationship of time since quitting to total mortality

 

Implement the test in Problem 14.49, and report a p-value (two-tailed)?


Health Promotion

A recent, article by Kenfield et al. [12] studied the relationship between various aspects of smoking and mortality among 104,519 women in the Nurses’ Health Study (NHS) from 1980−2004. One issue is whether there is a mortality benefit from quitting smoking vs. continuing to smoke and, if so, how long it takes for the mortality experience of former smokers to approximate that of never smokers. The data in Table 14.35 were presented comparing former smokers with current smokers.

Table 14.35: Relationship of time since quitting to total mortality

 

What test can be performed to compare mortality incidence between former smokers who quit <5 years ago vs. current smokers? 


Health Promotion

A recent, article by Kenfield et al. [12] studied the relationship between various aspects of smoking and mortality among 104,519 women in the Nurses’ Health Study (NHS) from 1980−2004. One issue is whether there is a mortality benefit from quitting smoking vs. continuing to smoke and, if so, how long it takes for the mortality experience of former smokers to approximate that of never smokers. The data in Table 14.35 were presented comparing former smokers with current smokers.

Table 14.35: Relationship of time since quitting to total mortality

 

What is the estimated mortality rate and 95% CI per 1000 person-years among current smokers? 


Health Promotion

A recent, article by Kenfield et al. [12] studied the relationship between various aspects of smoking and mortality among 104,519 women in the Nurses’ Health Study (NHS) from 1980−2004. One issue is whether there is a mortality benefit from quitting smoking vs. continuing to smoke and, if so, how long it takes for the mortality experience of former smokers to approximate that of never smokers. The data in Table 14.35 were presented comparing former smokers with current smokers.

Table 14.35: Relationship of time since quitting to total mortality

 

Suppose we look at the subset of women with a family history of colon cancer. Aspirin might be more beneficial in this high-risk subgroup. We have a total of 5000 person-years among ASA women and 2 events. We have a total of 20,000 person-years among control women and 20 events. Is there a significant difference in the incidence rates of colon cancer between these 2 groups? Provide a p-value (two-tailed).


Cancer

Suppose we wish to study the association between aspirin intake and the incidence of colon cancer. We find that 10% of women take 7 aspirin tablets per week (ASA group), while 50% of women never take aspirin (control group). The ASA group is followed for 50,000 person-years, during which 34 new colon cancers occurred over a 20-year period. The control group is followed for 250,000 person-years, during which 251 new colon cancers developed over a 20-year period.

Provide a 95% CI for the rate ratio in Problem 14.45?


Cancer

Suppose we wish to study the association between aspirin intake and the incidence of colon cancer. We find that 10% of women take 7 aspirin tablets per week (ASA group), while 50% of women never take aspirin (control group). The ASA group is followed for 50,000 person-years, during which 34 new colon cancers occurred over a 20-year period. The control group is followed for 250,000 person-years, during which 251 new colon cancers developed over a 20-year period.

What is the estimated rate ratio for colon cancer between the ASA and placebo groups?


Cancer

Suppose we wish to study the association between aspirin intake and the incidence of colon cancer. We find that 10% of women take 7 aspirin tablets per week (ASA group), while 50% of women never take aspirin (control group). The ASA group is followed for 50,000 person-years, during which 34 new colon cancers occurred over a 20-year period. The control group is followed for 250,000 person-years, during which 251 new colon cancers developed over a 20-year period.

Is there a significant difference between these incidence rates? Report a p-value (two-tailed)?


Cancer

Suppose we wish to study the association between aspirin intake and the incidence of colon cancer. We find that 10% of women take 7 aspirin tablets per week (ASA group), while 50% of women never take aspirin (control group). The ASA group is followed for 50,000 person-years, during which 34 new colon cancers occurred over a 20-year period. The control group is followed for 250,000 person-years, during which 251 new colon cancers developed over a 20-year period.

Answer Problem 14.40 under this definition of success?


To reduce variability, the investigators also considered a criterion of at least 50% improvement on two successive visits as a definition of success. 

Answer Problem 14.39 under this definition of success?


To reduce variability, the investigators also considered a criterion of at least 50% improvement on two successive visits as a definition of success. 

Divide participants according to median log10CO (adjusted), and estimate survival curves for each subgroup?

A person is selected randomly from the population and followed for 1.5 years. If the true rate of allergic reactions is 5 reactions per 100 person-years, what is the probability that the subject will have at least one allergic reaction during the follow-up period (i.e., cumulative incidence)? 


Infectious Disease

Suppose the rate of allergic reactions in a certain population is constant over time.

Two hundred subjects are selected randomly from the population and followed for various lengths of time. The average length of follow-up is 1.5 years. Suppose that at the end of the study, the estimated rate is 4 per 100 person- years. How many events must have been observed in order to yield the estimated rate of 4 per 100 person-years? 


Infectious Disease

Suppose the rate of allergic reactions in a certain population is constant over time.

Provide a 95% CI for the underlying rate, based on the observed data in Problem 14.23. Express the answer in units of number of events per 100 person-years?


Infectious Disease

Suppose the rate of allergic reactions in a certain population is constant over time.

Assess whether there is a difference between the incidence rate of breast cancer for premenopausal vs. postmenopausal women, while controlling for age. Report a p-value?


Cancer

The data in Table 14.34 provide the relationship between breast-cancer incidence rate and menopausal status by age, based on Nurses’ Health Study (NHS) data from 1976 to 1990.

Table 14.34 Relationship between breast-cancer incidence rate and menopausal status after controlling for age, NHS, 1976–1990

 



aPer 100,000 person-years.

Estimate the rate ratio for postmenopausal vs. premenopausal women after controlling for age. Provide a 95% CI for the rate ratio?


Cancer

The data in Table 14.34 provide the relationship between breast-cancer incidence rate and menopausal status by age, based on Nurses’ Health Study (NHS) data from 1976 to 1990.

Table 14.34 Relationship between breast-cancer incidence rate and menopausal status after controlling for age, NHS, 1976–1990

 



aPer 100,000 person-years.

1. Suppose a serum-creatinine level of ≥1.5 mg/dL is considered a sign of possible kidney toxicity. Use survivalanalysis methods to assess whether there are differences in the incidence of kidney toxicity between the high-N-acetylp- aminophenol (NAPAP) group and the control group. In this analysis, exclude subjects who were ≥1.5 mg/dL at baseline. 

2. Answer Problem 14.27 comparing the low-NAPAP group with the control group?

Answer Problem 14.27 while controlling for possible age and initial-level differences between groups. Consider both parametric and nonparametric survival analysis methods?


One issue is that the groups in Problem 14.27 may not be exactly balanced by age and/or initial level of serum creatinine. 

Assess the validity of the proportional-hazards assumption in Problems 14.29 and 14.30?


One issue is that the groups in Problem 14.27 may not be exactly balanced by age and/or initial level of serum creatinine.

A group of high-risk high school students was identified in the winter of 2008−2009 who had 3+ previous episodes of influenza before December 21, 2008. There were 20 students in this group, each followed for 90 days, of whom 8 developed influenza. Test the hypothesis that the high-risk students had a higher incidence rate of influenza than the average high school student during the winter of 2008−2009. Report a one-tailed p-value?


Infectious Disease

Suppose the incidence rate of influenza (flu) during the winter of 2008−2009 (i.e., from December 21, 2008, to March 20, 2009) was 50 events per 1000 person-months among students in high schools in a particular city.

Provide a two-sided 95% CI for incidence rate of flu among high-risk students during winter 2008−2009?


Infectious Disease

Suppose the incidence rate of influenza (flu) during the winter of 2008−2009 (i.e., from December 21, 2008, to March 20, 2009) was 50 events per 1000 person-months among students in high schools in a particular city.

What is the estimated incidence rate of flu in the 2009−2010 winter season per 1000 person-months? 


Among 1200 students in one high school in the city, 200 developed a new case of influenza over the 90 days from December 21, 2009, to March 20, 2010. 

Provide a 95% CI for the rate estimated in Problem 14.34?


Among 1200 students in one high school in the city, 200 developed a new case of influenza over the 90 days from December 21, 2009, to March 20, 2010. 

Test the hypothesis that the rate of flu has changed from the 2008−2009 to 2009−2010 winter season. Report a two-tailed p-value?


Among 1200 students in one high school in the city, 200 developed a new case of influenza over the 90 days from December 21, 2009, to March 20, 2010.

If the visual analog scale is treated as a continuous variable, assess whether there are any between-group differences in efficacy without considering the covariates. Try to do at least one analysis that uses the entire data set rather than focusing on specific time points?

Repeat the analysis in Problem 14.37, but account for covariate differences between groups?

Another way to score the visual analog scale is as a categorical variable where ≥50% improvement is considered a success and <50% improvement, remaining the same, or worsening is considered a failure. Answer the question posed in Problem 14.37 using the success/ failure scoring. Note that a patient may be a success at one visit but a failure at succeeding visits?

Repeat the analyses in Problem 14.39, but account for covariate differences between groups?

What are the estimated incidence rates in the ASA and control groups? 


Cancer

Suppose we wish to study the association between aspirin intake and the incidence of colon cancer. We find that 10% of women take 7 aspirin tablets per week (ASA group), while 50% of women never take aspirin (control group). The ASA group is followed for 50,000 person-years, during which 34 new colon cancers occurred over a 20-year period. The control group is followed for 250,000 person-years, during which 251 new colon cancers developed over a 20-year period.

What is a hazard rate in the context of this study? 


Pulmonary Disease

A study was performed among 169,871 Chinese men and women in 1991 ages 40 years and older [13]. Baseline data were collected in 1991, and a follow-up exam was conducted in 1999−2000. One component of the follow-up exam was a mortality follow-up for subjects who died between 1991 and 1999, where the date and cause of death were determined from Chinese vital statistics data. Of particular interest were risk factors for death from chronic-obstructive pulmonary disease (COPD), A Cox proportional-hazards regression model was used to relate risk factors in 1991 to time of death from COPD between 1991 and 1999−2000.

Write down the Cox proportional-hazards model. What does the term proportional hazards mean?


Pulmonary Disease

A study was performed among 169,871 Chinese men and women in 1991 ages 40 years and older [13]. Baseline data were collected in 1991, and a follow-up exam was conducted in 1999−2000. One component of the follow-up exam was a mortality follow-up for subjects who died between 1991 and 1999, where the date and cause of death were determined from Chinese vital statistics data. Of particular interest were risk factors for death from chronic-obstructive pulmonary disease (COPD), A Cox proportional-hazards regression model was used to relate risk factors in 1991 to time of death from COPD between 1991 and 1999−2000.

What is the hazard ratio for COPD mortality among men for smokers of ≥20 pack-years vs. never smokers? Provide a 95% CI.



The results in Table 14.37 were obtained from the study.

Table 14.37 Risk factors for COPD death among 169,871 study participants, China, 1991−2000

 

Test the hypothesis that the hazard ratio for smoking ≥20 pack-years vs. never smoking is significantly different (at the 5% level) for men vs. women?


Most risk factors seem of comparable magnitude for men and women. However, one exception is cigarette smoking. 

Test the hypothesis that the hazard ratio for smoking ≥20 pack-years vs. never smoking is significantly different (at the 5% level) for men vs. women?


Most risk factors seem of comparable magnitude for men and women. However, one exception is cigarette smoking. 

Suppose we wish to conduct a clinical trial to study the association between ASA and colon cancer. We will enroll 50,000 women in the study, half of whom will be assigned to ASA and half to placebo. Each woman is followed for 5 years. The expected incidence rate of colon cancer in the ASA group = 70/105 person-years and in the placebo group = 100/105 person-years. If we conduct a two-sided test with α = 0.05, how much power will the study have? 


Most risk factors seem of comparable magnitude for men and women. However, one exception is cigarette smoking. 

Suppose we want to enroll n subjects per group in the previously proposed study and follow each woman for 5 years. How many subjects do we need to achieve 90% power?


Most risk factors seem of comparable magnitude for men and women. However, one exception is cigarette smoking. 

Compare breast cancer incidence between the two exposure groups, where group 3 is the current PMH subjects and group 2 is the never PMH subjects?


Cancer

The data set in file BREAST.DAT consists of 1200 women in the NHS. The women were ascertained in 1990 and were postmenopausal and free of any cancer as of 1990. The 1200 women were selected in such a way that 200 of the women were current postmenopausal hormone (PMH) users in 1990 and 1000 of the women had never used PMH as of 1990. The objective of the analysis was to relate breast cancer incidence from 1990 to 2000 to PMH use as of 1990. Fifty-three of the women developed breast cancer between 1990 and 2000. PMH use is characterized both by current use/never use in 1990 as well as by duration of use as of 1990. Some current users in 1990 may have duration of use of 0 as of 1990 if they just started use in 1990 or if they used other types of PMH as of 1990 (other than estrogen or estrogen plus progesterone). There are two duration variables according to type of PMH use (duration of estrogen use in months as of 1990 and duration of estrogen plus progesterone use in months as of 1990). Each woman has a date of return of the 1990 questionnaire and a follow-up date = date of diagnosis of breast cancer if a case, or date of the last questionnaire filled out up to 2000 if a control. On the data file we provide the length of followup= follow-up date - date of return of the 1990 questionnaire (variable 18). Thus, the first subject (ID 10013) had a date of return of 1087 = July 1990 = (12 × 90 + 7) and a follow-up date of 1206 = June 2000 = (12 × 100 + 6). In addition, the file contains the values of other breast cancer risk factors as of 1990. A description of the data set is given in BREAST.DOC.


Format for Breast Cancer—postmenopausal hormone file

1. ID

2. Case 1=case, 0=control

3. Age

4. Age at menarche

5. Age at menopause

6. Age at first birth 98=nullip

7. Parity

8. Benign Breast disease (bbd) 1=yes/0=no

9. Family history of breast cancer 1=yes/0=no

10. BMI (kg/m**2)

11. Height (inches)

12. Alcohol use (grams/day)

13. PMH status 2=never user/3=current user

14. Duration of Estrogen use (months)

15. Duration of Estrogen + progesterone use (months)

16. Current Smoker 1=yes/0=no

17. Past smoker 1=yes/0=no

18. Follow-up time (months)

Compare the current PMH users vs. the never users in 1990 on other possible confounding variables?


Cancer

The data set in file BREAST.DAT consists of 1200 women in the NHS. The women were ascertained in 1990 and were postmenopausal and free of any cancer as of 1990. The 1200 women were selected in such a way that 200 of the women were current postmenopausal hormone (PMH) users in 1990 and 1000 of the women had never used PMH as of 1990. The objective of the analysis was to relate breast cancer incidence from 1990 to 2000 to PMH use as of 1990. Fifty-three of the women developed breast cancer between 1990 and 2000. PMH use is characterized both by current use/never use in 1990 as well as by duration of use as of 1990. Some current users in 1990 may have duration of use of 0 as of 1990 if they just started use in 1990 or if they used other types of PMH as of 1990 (other than estrogen or estrogen plus progesterone). There are two duration variables according to type of PMH use (duration of estrogen use in months as of 1990 and duration of estrogen plus progesterone use in months as of 1990). Each woman has a date of return of the 1990 questionnaire and a follow-up date = date of diagnosis of breast cancer if a case, or date of the last questionnaire filled out up to 2000 if a control. On the data file we provide the length of followup= follow-up date - date of return of the 1990 questionnaire (variable 18). Thus, the first subject (ID 10013) had a date of return of 1087 = July 1990 = (12 × 90 + 7) and a follow-up date of 1206 = June 2000 = (12 × 100 + 6). In addition, the file contains the values of other breast cancer risk factors as of 1990. A description of the data set is given in BREAST.DOC.


Format for Breast Cancer—postmenopausal hormone file

1. ID

2. Case 1=case, 0=control

3. Age

4. Age at menarche

5. Age at menopause

6. Age at first birth 98=nullip

7. Parity

8. Benign Breast disease (bbd) 1=yes/0=no

9. Family history of breast cancer 1=yes/0=no

10. BMI (kg/m**2)

11. Height (inches)

12. Alcohol use (grams/day)

13. PMH status 2=never user/3=current user

14. Duration of Estrogen use (months)

15. Duration of Estrogen + progesterone use (months)

16. Current Smoker 1=yes/0=no

17. Past smoker 1=yes/0=no

18. Follow-up time (months)

Perform an adjusted analysis comparing breast cancer incidence between the current PMH users vs. the never PMH users, adjusting for confounders that you found in Problem 14.65?


Cancer

The data set in file BREAST.DAT consists of 1200 women in the NHS. The women were ascertained in 1990 and were postmenopausal and free of any cancer as of 1990. The 1200 women were selected in such a way that 200 of the women were current postmenopausal hormone (PMH) users in 1990 and 1000 of the women had never used PMH as of 1990. The objective of the analysis was to relate breast cancer incidence from 1990 to 2000 to PMH use as of 1990. Fifty-three of the women developed breast cancer between 1990 and 2000. PMH use is characterized both by current use/never use in 1990 as well as by duration of use as of 1990. Some current users in 1990 may have duration of use of 0 as of 1990 if they just started use in 1990 or if they used other types of PMH as of 1990 (other than estrogen or estrogen plus progesterone). There are two duration variables according to type of PMH use (duration of estrogen use in months as of 1990 and duration of estrogen plus progesterone use in months as of 1990). Each woman has a date of return of the 1990 questionnaire and a follow-up date = date of diagnosis of breast cancer if a case, or date of the last questionnaire filled out up to 2000 if a control. On the data file we provide the length of followup= follow-up date - date of return of the 1990 questionnaire (variable 18). Thus, the first subject (ID 10013) had a date of return of 1087 = July 1990 = (12 × 90 + 7) and a follow-up date of 1206 = June 2000 = (12 × 100 + 6). In addition, the file contains the values of other breast cancer risk factors as of 1990. A description of the data set is given in BREAST.DOC.


Format for Breast Cancer—postmenopausal hormone file

1. ID

2. Case 1=case, 0=control

3. Age

4. Age at menarche

5. Age at menopause

6. Age at first birth 98=nullip

7. Parity

8. Benign Breast disease (bbd) 1=yes/0=no

9. Family history of breast cancer 1=yes/0=no

10. BMI (kg/m**2)

11. Height (inches)

12. Alcohol use (grams/day)

13. PMH status 2=never user/3=current user

14. Duration of Estrogen use (months)

15. Duration of Estrogen + progesterone use (months)

16. Current Smoker 1=yes/0=no

17. Past smoker 1=yes/0=no

18. Follow-up time (months)

Are there any interaction effects between PMH exposure and other risk factors?


Cancer

The data set in file BREAST.DAT consists of 1200 women in the NHS. The women were ascertained in 1990 and were postmenopausal and free of any cancer as of 1990. The 1200 women were selected in such a way that 200 of the women were current postmenopausal hormone (PMH) users in 1990 and 1000 of the women had never used PMH as of 1990. The objective of the analysis was to relate breast cancer incidence from 1990 to 2000 to PMH use as of 1990. Fifty-three of the women developed breast cancer between 1990 and 2000. PMH use is characterized both by current use/never use in 1990 as well as by duration of use as of 1990. Some current users in 1990 may have duration of use of 0 as of 1990 if they just started use in 1990 or if they used other types of PMH as of 1990 (other than estrogen or estrogen plus progesterone). There are two duration variables according to type of PMH use (duration of estrogen use in months as of 1990 and duration of estrogen plus progesterone use in months as of 1990). Each woman has a date of return of the 1990 questionnaire and a follow-up date = date of diagnosis of breast cancer if a case, or date of the last questionnaire filled out up to 2000 if a control. On the data file we provide the length of followup= follow-up date - date of return of the 1990 questionnaire (variable 18). Thus, the first subject (ID 10013) had a date of return of 1087 = July 1990 = (12 × 90 + 7) and a follow-up date of 1206 = June 2000 = (12 × 100 + 6). In addition, the file contains the values of other breast cancer risk factors as of 1990. A description of the data set is given in BREAST.DOC.

Assess whether the baseline level of visual field differs between RHO and RPGR patients?


Opththalmology

Retinitis pigmentosa (RP) is a hereditary ocular disease in which patches of pigment appear on the retina; the condition can result in substantial losses of vision and, in some cases, complete blindness. There are various modes of inheritance, including a dominant form, a recessive form, and a sex-linked form. An important discovery over the past 10 years was a set of genes that account for many of the RP cases. Specifically, mutations in the rhodopsin gene (RHO) account for many of the dominantly inherited cases; mutations in the RPGR gene (RPGR) account for many of the sex-linked cases. An important issue is whether the rate of progression is different between the RHO patients and the RPGR patients. On the data file FIELD.DAT are visual field data from approximately 100 patients in each group. Visual field is a measure of area of vision. It is measured in degrees2. Longitudinal data with varying follow-up times are provided for each patient separately for the right eye (labeled OD) and the left eye (labeled OS). Follow-up time varies from a minimum of 3 years to a maximum of about 25−30 years. 

Assess whether the rate of decline differs between RHO and RPGR patients?


Opththalmology

Retinitis pigmentosa (RP) is a hereditary ocular disease in which patches of pigment appear on the retina; the condition can result in substantial losses of vision and, in some cases, complete blindness. There are various modes of inheritance, including a dominant form, a recessive form, and a sex-linked form. An important discovery over the past 10 years was a set of genes that account for many of the RP cases. Specifically, mutations in the rhodopsin gene (RHO) account for many of the dominantly inherited cases; mutations in the RPGR gene (RPGR) account for many of the sex-linked cases. An important issue is whether the rate of progression is different between the RHO patients and the RPGR patients. On the data file FIELD.DAT are visual field data from approximately 100 patients in each group. Visual field is a measure of area of vision. It is measured in degrees2. Longitudinal data with varying follow-up times are provided for each patient separately for the right eye (labeled OD) and the left eye (labeled OS). Follow-up time varies from a minimum of 3 years to a maximum of about 25−30 years. 

Perform a test for trend based on the data in Table 14.41?


Cancer

A paper was recently published concerning the association between alcohol consumption and breast cancer incidence based on Nurses’ Health Study data from 1980–2006 (Chen et al. [16]). The following table was presented

Table 14.41: Association between alcohol intake in 1980 and breast cancer incidence from 1980 to 2006.

 


*per 105 person-years.

Provide a 95% CI for the IRR in Problem 14.82?


Cancer

A paper was recently published concerning the association between alcohol consumption and breast cancer incidence based on Nurses’ Health Study data from 1980–2006 (Chen et al. [16]). The following table was presented

Table 14.41: Association between alcohol intake in 1980 and breast cancer incidence from 1980 to 2006.

 


*per 105 person-years.

What is the estimated incidence rate ratio comparing the ≥30.0g group vs. the 0g group?


Cancer

A paper was recently published concerning the association between alcohol consumption and breast cancer incidence based on Nurses’ Health Study data from 1980–2006 (Chen et al. [16]). The following table was presented

Table 14.41: Association between alcohol intake in 1980 and breast cancer incidence from 1980 to 2006.

 


*per 105 person-years.

Perform the test in Problem 14.80, and report a p-value (two-tail)?


Cancer

A paper was recently published concerning the association between alcohol consumption and breast cancer incidence based on Nurses’ Health Study data from 1980–2006 (Chen et al. [16]). The following table was presented

Table 14.41: Association between alcohol intake in 1980 and breast cancer incidence from 1980 to 2006.

 


*per 105 person-years.

What test can be performed to compare the breast cancer incidence rate between the ≥ 30.0g group and the 0g group? 


Cancer

A paper was recently published concerning the association between alcohol consumption and breast cancer incidence based on Nurses’ Health Study data from 1980–2006 (Chen et al. [16]). The following table was presented

Table 14.41: Association between alcohol intake in 1980 and breast cancer incidence from 1980 to 2006.

 



*per 105 person-years.

Interpret what the results mean. Specifically, did the screening program have a significant impact on breast cancer mortality or not? 



One issue is that there may be time trends in breast cancer mortality that have nothing to do with screening. To control for this, a non-screening time period was identified (period C) during 1996–2005 in each region during which breast cancer screening was not introduced (in the above region it would be from 1996–2000). In addition, a time period 10 years prior to this non-screening period was identified during 1986–1995 (period D) (in the above region from 1986 1990). The breast cancer mortality exposure in these two periods was as follows: 

Table 14.40: Comparison of breast cancer mortality between period C and period D

 

Test the hypothesis that the IRR comparing period A to period B is the same as from period C to period D?



One issue is that there may be time trends in breast cancer mortality that have nothing to do with screening. To control for this, a non-screening time period was identified (period C) during 1996–2005 in each region during which breast cancer screening was not introduced (in the above region it would be from 1996–2000). In addition, a time period 10 years prior to this non-screening period was identified during 1986–1995 (period D) (in the above region from 1986 1990). The breast cancer mortality exposure in these two periods was as follows: 

Table 14.40: Comparison of breast cancer mortality between period C and period D

 

Estimate the incidence rate difference (IRD) comparing mortality incidence between period A and period B, and provide a 95% CI about this estimate?


Cancer

A study was conducted concerning the effect of screening mammography on breast cancer mortality in Norway (Kalager et al. [15]). A recommendation was made starting in 1996 to offer screening mammography to women ages 50–69. The 19 counties in Norway were grouped into 6 regions, and the screening project was introduced in a staggered fashion starting from 1996 until 2005. Thus, each region had a time period from part of 1996 to 2005 during which screening was offered (period A). In addition, historical information was used to estimate breast cancer mortality from 1986–1995 prior to when screening was introduced (period B). Thus, if screening was in place from 2001–2005 in a region, the historical exposure from 1991–1995 was considered in the same region. The results were as follows: 

Table 14.39 Comparison of breast cancer mortality between period A and period B

 

Estimate the incidence rate ratio (IRR) comparing mortality incidence between period A and period B, and provide a 95% CI about this estimate?


Cancer

A study was conducted concerning the effect of screening mammography on breast cancer mortality in Norway (Kalager et al. [15]). A recommendation was made starting in 1996 to offer screening mammography to women ages 50–69. The 19 counties in Norway were grouped into 6 regions, and the screening project was introduced in a staggered fashion starting from 1996 until 2005. Thus, each region had a time period from part of 1996 to 2005 during which screening was offered (period A). In addition, historical information was used to estimate breast cancer mortality from 1986–1995 prior to when screening was introduced (period B). Thus, if screening was in place from 2001–2005 in a region, the historical exposure from 1991–1995 was considered in the same region. The results were as follows: 

Table 14.39 Comparison of breast cancer mortality between period A and period B

 

There were 722 patients admitted to the ICU for H1N1 in Australia and New Zealand from June 1 to August 31, 2009 (winter season). The total population of Australia and New Zealand is approximately 25,000,000 = 25 million. Suppose that the underlying incidence rate is the same for the United States as for Australia and New Zealand. If there are 250 million people who live in the United States, then what is the estimated number of H1N1 cases admitted to the ICU in the United States during the winter season (12/21/09−3/20/10), and what is a 95% CI for the number of US cases? 


The data in Table 14.38 were presented on H1N1 patients admitted to the ICU (no. per million inhabitants) by week and region. 

Table 14.38 H1N1 patients admitted to the ICU in 2009 (no. per million inhabitants) by week and region

 

Using the incidence rates calculated in Problem 14.73, test whether there are significant differences in incidence rates by region over the 7-week period. Provide a two-sided p-value. (Note: Assume that the underlying incidence rate in a given region is the same over the 7 weeks?)


The data in Table 14.38 were presented on H1N1 patients admitted to the ICU (no. per million inhabitants) by week and region. 

Table 14.38 H1N1 patients admitted to the ICU in 2009 (no. per million inhabitants) by week and region

 

For each region,

(i) Estimate the incidence rate per million (106) inhabitants (per week) over the 7-week period, and

(ii) The number of cases over 7 weeks from 06/29/09 to 08/16/09 (rounded to the nearest integer). 


The data in Table 14.38 were presented on H1N1 patients admitted to the ICU (no. per million inhabitants) by week and region. 

Table 14.38 H1N1 patients admitted to the ICU in 2009 (no. per million inhabitants) by week and region

 

It was found that among 626 Australian H1N1 patients admitted to an intensive care unit (ICU), 61 were aboriginal. If 2.5% of Australians are aboriginal, test the hypothesis that the percentage of aboriginal H1N1 cases differs from the percentage of aboriginals in the general population. Provide a two-sided p-value?


Infectious Disease

A study was recently performed concerning the incidence of H1N1 influenza in Australia and New Zealand [14].

Assess whether the time to legal blindness is the same or different between the RHO group and the RPGR group. For simplicity, assume that legal blindness in a particular eye is an absorbing state (i.e., once they become legally blind in an eye, they remain legally blind). Also, do not include eyes that are legally blind at baseline in the analysis, since they have already reached the endpoint. The format of the data set is given in FIELD.DOC?



An important endpoint for RP patients is legal blindness. For visual field, legal blindness is usually defined as <20° diameter of equivalent circular field area in the better eye. The equivalent circular field area for a 20° diameter is πR2 = π(10°)2 = 314 degrees squared. For example, ID 156 reached this endpoint at the fifth visit (approximately at 5.06 years of follow-up) in the right eye and at both eyes at the sixth visit (approximately 6 years of follow-up). 

Answer the question posed in Problem 14.69, while accounting for possible age and gender differences between groups?


FIELD.DOC

Column _________________________________________ Variable

1−6 ……………………………….............………………………..… ID

8 …………………………………... Group (1 = RHO/2 = RPGR)

11−14 …………………………..…. Age at visit (XX.X in years)

16 ………………....………… Gender (1 = m/2 = f) Gender is 

coded as missing, but is actually male for all members of the RPGR group.

18−27 …………..……………. Date of visit (month/day/year)

29−34 …….…....………………… Time from 1st visit in years

36−43 …..…. Total field area right eye (OD) in degrees2

45−52 …..…….. Total field area left eye (OS) in degrees2


One possible complexity is that the age distribution of the 2 groups may not be balanced. Also, the RHO group contains both males and females while the RPGR group consists exclusively of males. Gender is also missing for some subjects in the RHO group. Gender is coded as missing for subjects in the RPGR group because they are all male.

Is there a difference in incidence according to duration of use (Variables 14 and 15)?


Cancer

The data set in file BREAST.DAT consists of 1200 women in the NHS. The women were ascertained in 1990 and were postmenopausal and free of any cancer as of 1990. The 1200 women were selected in such a way that 200 of the women were current postmenopausal hormone (PMH) users in 1990 and 1000 of the women had never used PMH as of 1990. The objective of the analysis was to relate breast cancer incidence from 1990 to 2000 to PMH use as of 1990. Fifty-three of the women developed breast cancer between 1990 and 2000. PMH use is characterized both by current use/never use in 1990 as well as by duration of use as of 1990. Some current users in 1990 may have duration of use of 0 as of 1990 if they just started use in 1990 or if they used other types of PMH as of 1990 (other than estrogen or estrogen plus progesterone). There are two duration variables according to type of PMH use (duration of estrogen use in months as of 1990 and duration of estrogen plus progesterone use in months as of 1990). Each woman has a date of return of the 1990 questionnaire and a follow-up date = date of diagnosis of breast cancer if a case, or date of the last questionnaire filled out up to 2000 if a control. On the data file we provide the length of followup= follow-up date - date of return of the 1990 questionnaire (variable 18). Thus, the first subject (ID 10013) had a date of return of 1087 = July 1990 = (12 × 90 + 7) and a follow-up date of 1206 = June 2000 = (12 × 100 + 6). In addition, the file contains the values of other breast cancer risk factors as of 1990. A description of the data set is given in BREAST.DOC.

Showing 1 - 100 of 1422
Join SolutionInn Study Help for
1 Million+ Textbook Solutions
Learn the step-by-step answers to your textbook problems, just enter our Solution Library containing more than 1 Million+ textbooks solutions and help guides from over 1300 courses.
24/7 Online Tutors
Tune up your concepts by asking our tutors any time around the clock and get prompt responses.