Begun in 1985, this is a prospective cohort study of working men and women conducted at eight New York City work sites each employing at least 150 men. These sites are: a newspaper typography department, a federal health agency, a stock brokerage firm, a liquor marketer, a private hospital, a sanitation collection and repair facility, a department store warehouse and the headquarters for a large insurance company.
The methods for the first round of data collection of this study (Time 1) have been reported in detail elsewhere (JAMA 1990, JAMA 1992, Hypertension 1992). In brief, potential subjects received casual blood pressure screening by department at their worksites and demographic data was collected. For the employees of a department to be eligible for the study at least 75% had to participate in the screening.
From this screened sample, subjects were eligible to be selected for the study if they were between 30 and 60 years old, were employed more than 30 hours/week, were able to read English, had a body mass index (kg/m2) less than 32.5, had no second job of 15 or more hours per week and had been at their current worksite for at least three years before being approached for this study and, if applicable, before being diagnosed as having high blood pressure. To increase the number of eligible subjects for the cohort study at the eighth and last site the following eligibility criteria were changed: (1) subjects were no longer recruited in a fixed ratio (2:3) of cases to controls as at the previous sites; (2) subjects no longer needed to be at their worksite for three years but instead for only one year; and (3) those subjects who had casual screening and recruitment BP changes that resulted in their being classified as "crossovers" and excluded from the analysis for the initial Time 1 of the study were allowed to remain in the cohort study. The third criteria was no longer relevant to the outcome of interest (change in AmBP over time) None of these changes should effect either the effect estimates nor the validity of the findings.
Subjects were excluded from the study if they had a history of coronary, cerebrovascular, or peripheral vascular disease, electro-cardiographic evidence of myocardial infarction, ischemia or atrial fibrillation, funduscopic changes, evidence of any secondary cause of hypertension, screening systolic blood pressure greater than 160 mm Hg or screening diastolic blood pressure greater than 105 mm Hg. Subjects who reported any of the above mentioned cardiovascular events between Time 1 and Time 2 evaluation were excluded from these analyses.
Of the initial 285 male subjects recruited at Time 1, 195 subjects
were alive, located and completed Time 2. Twenty-three subjects
on anti-hypertensive medications at the time of evaluation at
Time 1 and/or Time 2 were included in most analyses despite the
potential influence of medications on AmBP changes over time.
Subjects wore an AmBP monitor (Spacelabs 5200) for 24 hours during a normal work day, using procedures described previously (James etal 1986) at Time 1 and a Spacelabs 90202 device at Time 2 (ref needed). Readings were collected every 20 minutes during the day at Time 2 instead of every 15 minutes at Time 1. Otherwise the methodology of AmBP data collection was unchanged. The monitor was attached at either the subject's work site or at the Hypertension Center at Cornell University Medical College (CUMC) Hypertension Center and calibrated by comparing five successive systolic and diastolic readings against simultaneously determined auscultatory readings taken by a trained observer with a mercury column, in which both had to be within 5 mm Hg to be acceptable. The timer on the monitor was set to take readings at 20 minute intervals during the day, and 30 minute intervals during normal hours of sleep, and the subject was instructed to proceed through a normal workday. Each time the monitor inflated and recorded blood pressure during waking hours the subject was asked to remain as motionless as possible and then to record his activity, location, position, and mood in a diary. The diary information (i.e., whether subjects reported being at work, home or sleep) was used to calculate average AmBP for each location category.
Subjects also received a routine medical examination, which included a full history, a cardiovascular physical examination, assessment of alcohol intake, current smoking history, and exercise habits. Height and weight were determined at the physical examination and body mass index (BMI) was calculated according to the formula - weight(kg)/height(m)2. Blood testing, EKG and an M-mode echocardiogram under two dimensional guidance performed according to standard procedures (Devereaux etal 1986) were carried out at the Hypertension Center at CUMC.
Subjects completed a questionnaire packet which included the Job Content Questionnaire (JCQ) to evaluate 'job strain'. The JCQ is a 42-item questionnaire developed by Dr. Robert Karasek, based, in part, on questions drawn from the US. Department of Labor/University of Michigan Quality of Employment Surveys (Karasek etal 1985). Two scales were used to define 'job strain' - job decision latitude and psychological job demands (see Fig. 1). Job decision latitude, an operationalization of the concept of "job control" was defined as the sum of two subscales each given equal weight: 1) skill discretion, measured by 6 items (keep learning new things; can develop skills; job requires skill; task variety; repetitious; job requires creativity), and 2) decision authority, measured by 3 items (have freedom to make decisions; can choose how to perform work; have a lot of say on the job). Psychological job demands was defined by 5 items (excessive work; conflicting demands; insufficient time to do work; work fast; work hard). All questions are scored on a Likert scale of 1 to 4, and both decision latitude and psychological job demands were constructed to have a range of 12 to 48. Scale reliability was acceptable for both decision latitude (Cronbach's alpha=.82) and workload demands (alpha=.74).
Previous research (Karasek etal 1988) in a nationally representative working male population indicated that about 20% of the men have jobs simultaneously high in demands and low in decision latitude, a situation labeled 'job strain' or "high strain jobs". Cut points for psychological workload demands and job decision latitude were selected so that 20% of our study sample would also be classified as having 'high strain' jobs. Jobs were classified as 'high strain' if subjects scored both 37 or below for decision latitude and 32 or above for psychological job demands. In addition, the jobs of persons located in the three other quadrants defined by these cut points were labeled active, passive and low strain as shown in Figure 1 (e.g., active jobs were those in which subjects scored both 37 or above for decision latitude and 32 or above for psychological job demands, etc.).
For the cohort analysis a new categorical job strain variable was constructed based on the subjects 'job strain' scores at Time 1 and Time 2. Subjects reporting no 'job strain' at both Time 1 and Time 2 (Group 1, N=138) are the referent group. Subjects without job strain at Time 1 and with 'job strain' at Time 2 are Group 2, N=17, while those reporting 'job strain' at Time 1 but no 'job strain' at Time 2 are Group 3, N=25. Subjects who reported 'job strain' at both Time 1 and Time 2 are classified as having 'chronic job strain' Group 4, N=15 (see Fig. 1).
The same battery of psychosocial questionnaires as were administered
at Time 1 were again administered at Time 2. The Jenkins Activity
Survey was administered to evaluate Type A behavior, and subjects
were classified as 'Type A' if they scored above 0. A demographic
questionnaire elicited information on years of education, individual
and family income, marital status, religion, race, age, and employment
history. Age was treated as a continuous variable in this analysis.
Education was included in this analysis as a control variable
because of the known potential impact of low socioeconomic status
(often measured as low education) on blood pressure and was entered
as a continuous variable in years of education when used as a
control variable or as a dichotomous variable for analyses examining
effects on AmBP (1 for subjects with 12 years or less education,
0 for subjects with >12 years of education). Alcohol and smoking
behavior were assessed by questionnaire at the time of the medical
examination with the responses reviewed by a nurse. Subjects were
classified either as non-drinkers if they reported they drink
not at all or occasionally, or as drinkers if they reported regular
consumption (at least 4 or more times per week) or binge drinking.
Subjects were classified as smokers if they currently smoke. Race
was classified as either Caucasian or other. Finally, physical
activity on the job was evaluated by a single item from the JCQ
('job requires lots of physical effort') and scored on a Likert
scale of 1 to 4. "Job title change" was assessed by
an interviewer and defined as either change in job title or a
significant change in job duties even if a subjects' job title
was officially unchanged.
A total of 90 subjects who participated at Time 1 were excluded from analysis at Time 2 of the study for the following reasons: 3 were deceased, 6 developed cardiovascular disease, 16 were not employed due to retirement, unemployment, 6 could not be located, 41 refused participation, and 18 failed to complete the protocol (see Table 1 where N=65, those deceased or have developed cardiovascular disease N=9, and those not employed N=16 are excluded due to lack of eligibility for the 2nd wave of study). There are very little missing data for the subjects included in this analysis. Those missing data on either the outcome measure, AmBP or 'job strain', the focal predictor, were excluded from the analyses of that outcome. The modal category (or mean) has been substituted for missing data on all categorical (or continuous) covariates.