Possible effect of DRGs on the classification of stroke: implications for epidemiological surveillance.

BACKGROUND AND PURPOSE: Accurate data on the distribution of stroke subtypes are essential for understanding the forces driving recent morbidity and mortality trends. The introduction of diagnosis-related groups (DRGs) in the 1980s may have affected the distribution of stroke subtypes as defined by International Classification of Diseases, Ninth Revision (ICD-9), discharge diagnosis codes. METHODS: The Pawtucket Heart Health Program cardiovascular surveillance data were used to examine trends in stroke classification for 1980 to 1991 in relation to the introduction of DRGs in 2 communities in Massachusetts and Rhode Island, where DRGs were implemented 2 years apart. Included were all hospital discharges for residents aged 35 to 74 with a primary ICD-9 diagnosis of 431 to 432, 434, or 436 to 437 (N=1386 in Rhode Island, N=1839 in Massachusetts). RESULTS: In each state, concurrently with the introduction of DRGs, the proportion of strokes classified as cerebral occlusion (ICD-9 434.0 to 434.9) increased, and the proportion classified as acute but ill-defined (ICD-9 436.0 to 436.9) decreased. Before DRGs, 30.0% of strokes in Rhode Island and 26.6% in Massachusetts were classified as cerebral occlusion, whereas 51.8% in Rhode Island and 51.7% in Massachusetts were classified as acute ill defined. After DRGs were instituted, the proportions of cerebral occlusion and acute, ill-defined stroke, respectively, were 70.9% and 8.5% in Rhode Island and 74.1% and 7.7% in Massachusetts (chi(2), all P<0.001). The proportions of strokes classified as intracerebral hemorrhage or transient cerebral ischemia remained constant. CONCLUSIONS: The implementation of DRGs may have influenced coding of strokes to the ICD-9. Findings highlight the limitations of hospital discharge data for evaluating stroke subtypes and demonstrate the need for community-based surveillance for monitoring specific trends in stroke.

Coronary heart disease surveillance: field application of an epidemiologic algorithm.

This report describes the performance of a surveillance system and computerized algorithm for the assignment of definite or probable hospitalized cardiac events for large epidemiologic studies. The algorithm, developed by the Coordinating Committee for Community Demonstration Studies (CCCDS), evolved from the Gillum criteria, and included selected ICD-9-CM codes including codes 410 through 414 for discharge record screening, plus creatine kinase. For the small percentage of cases in which enzyme analysis was inconclusive (8%), presence of pain and/or Minnesota-coded electrocardiograms were included to define the outcome. All data items were easily obtained from medical records by trained lay record abstractors and required no interpretation. From January 1980 through December 1991, 21,183 medical records were screened for ICD-9-CM codes 410 through 414. Of all 410 to 411 ICD-9-CM codes (n = 9026), 36.9% (n = 3220) were classified as definite cardiac events and 10.6% (n = 1057) as probable events. Of all 412 through 414 codes (n = 9070), only 1.8% (n = 227) were classified as definite cardiac events and 5.4% (n = 716) as probable events. The epidemiologic diagnostic algorithm presented in this article used computerized data to assign diagnoses in a standard, objective manner, and was a lower cost alternative to classification of cardiac events on the basis of clinical review and/or more complex record abstraction approaches.

Sex-specific trends in validated coronary heart disease rates in southeastern New England, 1980-1991.

Although the national decline in coronary heart disease mortality began earlier and was steeper in women relative to men, recent data suggest that the decline in women has slowed. The purpose of this study was to document sex-specific trends in coronary disease morbidity and mortality for the period 1980-1991 in two southeastern New England communities, and to determine whether temporal trends have been similar in men and women aged 35-74 years. Analyses were based on 6,282 validated in-hospital and out-of-hospital coronary disease events ascertained by the retrospective surveillance system of the Pawtucket Heart Health Program. Total (fatal plus non-fatal) coronary disease rates remained stable during this period. The flat trend was the result of an increase in non-fatal hospitalizations and a simultaneous decrease in both in-hospital and out-of-hospital mortality. The decline in fatal coronary disease was steeper for men, for both in- and out-of-hospital mortality, although the sex difference was statistically significant only for out-of-hospital deaths. In-hospital case-fatality for validated coronary disease declined for both men and women. The steeper decline in coronary disease mortality for men suggests the need for more information regarding sex differential trends in prevention, diagnosis, classification, and treatment.

Trends in validated cases of fatal and nonfatal stroke, stroke classification, and risk factors in southeastern New England, 1980 to 1991 : data from the Pawtucket Heart Health Program.

BACKGROUND AND PURPOSE: Recent US data suggest there is a slowing of the decline in stroke mortality rates, accompanied by a constant morbidity rate. Hospital discharge rates for patients with stroke are influenced by numerous factors, and community-based surveillance data for validated cases are rare. Thus, reasons for the observed trends remain unclear. In the present study, we examined trends in validated cases of stroke for 1980 to 1991 in the combined populations of the Pawtucket Heart Health Program study communities and examined concomitant trends in classification, use of diagnostic procedures, and levels of risk factors. METHODS: Discharges for residents aged 35 to 74 years with International Classification of Diseases, Ninth Revision codes 431, 432, and 434 to 437 were identified through retrospective surveillance. A physician reviewed the medical records to validate case status. RESULTS: Between 1980 and 1991, 2269 discharges were confirmed as representing definite or probable strokes (59.5% of 3811 cases reviewed). The fatal stroke rate declined (P<0.005) and the nonfatal stroke rate remained constant in both sexes. Case-fatality rates declined significantly (P=0.003), and among strokes, the relative odds of death in 1990 versus 1980 was 0.50 (95% CI 0.34 to 0.72). The proportion of stroke discharges in which the patient received a CT scan or MRI increased 120%, and fewer strokes were classified as ill defined. Hypertension prevalence, treatment, and control rates remained constant in these populations. CONCLUSIONS: Although causes for the observed trends remain unclear, results suggest that the decline in mortality rates is due to improved survival rates for patients with stroke. However, constant morbidity rates combined with constant rates of hypertension highlight the need for improved prevention to reduce the impact of stroke.

Effects of a community-based intervention on physical activity: the Pawtucket Heart Health Program.

OBJECTIVES: The effect of a community-based physical activity program in Pawtucket, RI, was evaluated relative to one in a comparison community. METHODS: Cross-sectional surveys of 7529 residents of Pawtucket, RI, and 7732 residents of the comparison city were conducted at 2-year intervals during 7 years of intervention. RESULTS: There were no differences in self-reported knowledge of the benefits of physical activity, attempts to increase exercise, or prevalence of physical inactivity between Pawtucket and the comparison community. CONCLUSIONS: Future community-based physical activity interventions should attempt to involve a wider range of individuals.

HDL cholesterol: trends in two southeastern New England communities, 1981-1993.

PURPOSE: Although public health interventions have not specifically targeted high density lipoprotein (HDL) cholesterol, observed changes in the prevalence of other cardiovascular risk factors would be expected to have differential effects on HDL. This study examined secular trends in HDL in relation to changes in other cardiovascular risk factors for the years 1981 through 1993 in the Pawtucket Heart Health Program (PHHP) study communities. METHODS: Nonfasting HDL levels were assessed in 12,223 respondents to six biennial population random sample surveys. RESULTS: Between 1981 and 1993, mean HDL cholesterol declined by 0.08 mmol/L in both men and women after adjustment for age, city, education, hormone use, medications, recent alcohol use, smoking, regular exercise, body mass index (BMI), and total cholesterol, (p for trend < 0.001). There was no apparent laboratory explanation for the trend which occurred concurrent with decreased smoking prevalence, increasing BMI and decreased prevalence of recent alcohol use. Decreasing HDL cholesterol was observed consistently across subgroups defined by smoking, alcohol use and BMI. CONCLUSIONS: Although several favorable cardiovascular risk factor trends have been observed in recent decades, declining HDL cholesterol is also of interest, particularly in conjunction with population increases in BMI.

Short stature and cardiovascular disease among men and women from two southeastern New England communities.

BACKGROUND: Short stature has been associated with an increased risk of coronary heart disease (CHD), although the reason for the association remains unclear. Data on the relation between stature and stroke is more limited. We examined the association between stature and CHD as well as between stature and stroke in men and women from two communities in southeastern New England. METHODS: Coronary heart disease and stroke events were abstracted from medical records between January 1980 and December 1991. An epidemiological diagnostic algorithm developed to measure CHD was used in the present analysis. Unadjusted relative risks (RR) and RR adjusted for age, smoking status, obesity, high-density lipoprotein (HDL) cholesterol <0.91 mmol/l, total cholesterol >6.21 mmol/l, hypertension, diabetes, education, and being foreign born were computed by gender-specific height categories separately for men (n = 2826) and women (n = 3741). RESULTS: A graded inverse association between stature and risk of CHD was observed among men which persisted after adjustment for confounders. Men >69.75 inches had an 83% lower risk of CHD compared with men < or = 65 inches. In addition, the tallest men had a 67% decreased risk of stroke compared with the shortest men. No significant relation between stature and CHD or stroke was observed among women. CONCLUSIONS: These data support the hypothesis that stature is inversely related to both risk of CHD and stroke at least among men. Factors which might explain this association remain to be determined.

Dietary factors in relation to weight change among men and women from two southeastern New England communities.

OBJECTIVE: Dietary factors, independent of total energy, may be important in promoting obesity. We examined prospectively the relationship between baseline diet composition and weight gain among adult men and women from southeastern New England. DESIGN: The prospective association of nutrient consumption and weight change was examined in a randomly selected cohort examined four years apart. SUBJECTS: Adults aged 18 through 64 years from two communities in Southeastern New England were randomly selected for the study after being interviewed in their homes. The present investigation is based on a subgroup of 465 individuals who completed a food-frequency questionnaire in 1986 or 1987 and were reinterviewed four years later. STATISTICAL ANALYSES: Multiple regression analyses were used to determine the association of weight change with different nutrients and food groups after adjusting for age, smoking status, baseline body mass index, physical activity level, and total energy. RESULTS: Total energy was positively associated with weight gain and age was inversely associated with weight gain. None of the nutrients or food groups were significantly related to weight gain. CONCLUSIONS: These findings indicate that weight gain increased with increasing baseline total energy intake, particularly in the young. Future research is required to determine ways of decreasing energy intake in younger individuals.

Is cigarette smoking an effect modifier of the relation between antihypertensive therapy and blood lipids?

We examined the relation between current use of antihypertensive medications and lipid levels in relation to smoking status. We used data from cross-sectional random sample surveys conducted between 1981 and 1993 in two southeastern New England communities. The analysis included 3,027 normotensive, 1,416 untreated hypertensive, and 1,174 treated hypertensive survey respondents between 40 and 64 years of age. After adjustment for body mass index, diabetes, mellitus, alcohol use, use of medications with adverse lipid effects, age, and sex, the estimated mean high-density-lipoprotein-cholesterol level among hypertensives using beta-blockers [44.6 mg per dl; 95% confidence interval (CI) = 43.1-46.1] was lower than that for either hypertensives (48.5 mg per dl; 95% CI = 47.8-49.2) or normotensives (47.1 mg per dl; 95% CI = 46.6-47.6). This pattern was similar among smokers and nonsmokers. Smokers treated with beta-blockers, however, had particularly low levels of high-density-lipoprotein-cholesterol (43.4 mg per dl; 95% CI = 40.7-46.1), compared with non-smokers using beta-blockers (45.8 mg per dl; 95% CI = 44.2-47.3). Low levels of high-density-lipoprotein-cholesterol may increase the risk of coronary heart disease among smokers using beta-blockers.

Correlates of oral contraceptive use in two New England communities: 1981-1993.

This study compared the sociodemographic and cardiovascular correlates of oral contraceptive users and nonusers between 1981 and 1993. We also sought to determine changes in the prevalence of oral contraceptive use among older premenopausal women. Using data from six biennial cross-sectional household surveys in a population-based sample of two New England communities, 5239 women between ages 18 and 45 years were identified. Women using oral contraceptives were 5 years younger and better educated than nonusers, and users more frequently reported a per capita income above the median for the survey population than nonusers. Users were more likely to report smoking and have a lower body mass index than nonusers. Little use of oral contraceptives was detected among women ages 40-45 years despite changes in FDA recommendations about their use and increasing evidence of their noncontraceptive benefits in healthy older nonsmoking women.

Self-reported alcohol intake using two different question formats in southeastern New England.

BACKGROUND: Quantitative measures of alcohol intake are not always available in population studies. METHOD: The authors evaluated whether a question on alcohol intake embedded within a general health survey could be used as a surrogate marker for alcohol intake. We compared alcohol intake assessed with a food frequency questionnaire (FFQ) with alcohol intake assessed with a simple dichotomous survey question in a population-based sample. RESULTS: The study population consisted of 755 men and 1164 women from two communities in southeastern New England in the USA. There was strong agreement between the two alcohol questions for the classification of nondrinkers (98.1%). When participants were classified according to the quantity of alcohol consumed on the FFQ, the ability of the simple question to identify drinkers improved in a dose-response fashion. The Kappa statistic was 0.08 (P < 0.001), 0.38 (P < 0.001), and 0.81 (P < 0.001) for low, medium, and high consumers of alcohol, respectively. CONCLUSIONS: These results suggest that the survey alcohol question provides a useful qualitative measure of categorizing nondrinkers and identifying drinkers who consume more than one drink per day. In population studies where quantitative measures of alcohol intake may not be available a survey alcohol question may prove useful when alcohol intake is likely to confound results, and adjustment of the data is needed.

High-density-lipoprotein cholesterol and types of alcoholic beverages consumed among men and women.

OBJECTIVES: Differences by sex in the relationship between high-density-lipoprotein (HDL) cholesterol and consumption of alcoholic beverages were examined in 1516 individuals. METHODS: Questionnaires and blood-sample data from cross-sectional surveys were analyzed. RESULTS: Both beer and liquor were independently associated with increased HDL cholesterol in the total group, in men, and in women after covariates were controlled for. Wine was associated with a significant increase in HDL cholesterol in women only. CONCLUSIONS: Among women and men, amount may be more important than type of alcoholic beverage consumed. The independent effect of wine on HDL cholesterol among men remains unclear since few men in this population consumed wine exclusively or in large quantities.

The association between perceived risk and actual cardiovascular disease.

We examined the association between self-rated risk of stroke or myocardial infarction and actual morbidity and mortality by gender among participants of the Pawtucket Heart Health Program Health Survey in a nested case-control study. We defined cases (N = 191) as survey participants who later had a cardiac event registered in the Pawtucket Heart Health Program cardiovascular disease morbidity and mortality surveillance system. We matched controls (N = 573) on age, sex, date of survey, and city of residence. Women reporting a high perceived risk of stroke or myocardial infarction were four times as likely as women who indicated low perceived risk to have a future cardiovascular event [adjusted odds ratio = 4.0; 95% confidence interval (CI) = 1.5-10.3]. Men who perceived their risk as average were twice as likely as those who rated their risk as low to have a future myocardial infarction or stroke (adjusted odds ratio = 1.9; 95% CI = 1.0-3.6), whereas those who perceived their risk as high were at only slightly higher risk of morbid or mortal events (adjusted odds ratio = 1.8; 95% CI = 0.9-3.9).

Sedentary lifestyle and risk of coronary heart disease in women.

Relatively few studies have assessed the prospective relationship of sedentary lifestyle and coronary heart disease (CHD) in women. We performed a nested case-control study, identifying 50 cases of acute CHD in women and 150 age-matched controls. Risk factors including sedentary lifestyle, cigarette smoking, hypertension, diabetes mellitus, total, and HDL cholesterol levels had been previously determined in this cohort using household survey techniques. The odds ratio (OR) of CHD in sedentary women was 2.1 (95% CI 1.0-4.3, P = 0.046). The odds ratio was similar (OR = 2.3, 95% CI 1.0-5.7, P = 0.051) after adjusting for potential confounders of hypertension, cigarette smoking, low HDL cholesterol, diabetes mellitus, and being U.S. born. This study suggests that sedentary lifestyle may be an independent risk factor for CHD in women, but its results are open to several contradictory interpretations. Further studies to investigate the prospective relationship of sedentary lifestyle and the risk of CHD in women are needed.

Cardiovascular risk profiles of oral contraceptive users and nonusers: a population-based study.

BACKGROUND: The purpose of this population-based study was to evaluate whether women taking oral contraceptive products (OCPs) have more favorable cardiovascular disease (CVD) risk profiles than women not taking OCPs. METHODS: Data including sociodemographic variables and CVD risk factors from 5,239 women of reproductive age were gathered in six cross-sectional household surveys. Differences in CVD risk factor profiles between OCP users and nonusers were compared using logistic regression. younger users and older users were compared with nonusers of the same age groups. RESULTS: Women using OCPs (n = 807) were younger and had higher socioeconomic status than nonusers (n = 4,432). More users than nonusers were current smokers (43.1% vs 36.6%, P = 0.0011), had a total serum cholesterol level > or = 240 mg/dl (12.7% vs 6.0%, P = 0.0001), and tended to rate themselves to be at low risk for CVD (48.3% vs 45.5%, P = 0.17). Fewer OCP users than nonusers had an HDL cholesterol level < or = 35 mg/dl (4.0% vs 6.9%, P = 0.0039). CONCLUSIONS: Although they tended to perceive themselves to be at low risk, women taking OCPs in this study had worse CVD risk profiles than women not reporting hormonal contraception and are candidates for cardiovascular risk reduction strategies.

Prior and current health characteristics of postmenopausal estrogen replacement therapy users compared with nonusers.

OBJECTIVE: Our purpose was to investigate whether selection of healthy women for postmenopausal estrogen therapy may confound observational studies of estrogen use and cardiovascular disease risk. STUDY DESIGN: Data were obtained from baseline (1981 to 1984) and follow-up (1990 to 1992) health surveys of two cohorts randomly selected from communities in southeastern New England. At follow-up postmenopausal women > or = 40 years old were categorized as current users (n = 70) or nonusers (n = 772) of noncontraceptive estrogen. Users and nonusers were compared on both prior characteristics from the baseline surveys and current characteristics measured at follow-up by use of analysis of covariance. RESULTS: Prior levels of total and high-density lipoprotein cholesterol, body mass index, and blood pressure were similar for estrogen users and nonusers. Estrogen users were less likely to have smoked and more likely to have had their cholesterol checked and to exercise regularly. These differences were more pronounced for current characteristics than for baseline characteristics. CONCLUSIONS: Selection of healthy women for treatment may not fully explain the apparent protective effect of estrogen replacement on cardiovascular risk. However, more healthy profiles among estrogen users may inflate the apparent benefit of treatment in observational studies.

Cross-sectional relationship between diet and physical activity in two southeastern New England communities.

The hypothesis that health promotive diets associated with higher levels of habitual physical activity confound the relationship between regular physical activity and health has not been well explored in epidemiologic studies. We evaluated self-reported physical activity, Willett Food Frequency dietary data, sociodemographic and physiologic factors cross-sectionally for 2,004 household survey participants in two southeastern New England communities. We compared the dietary habits of sedentary participants (n = 964) to those of moderately active (n = 600) and very active (n = 440) participants after adjusting for age, gender, education, smoking status, year of survey, and total calories. Our results showed that moderately active and very active participants consumed more fiber, less total fat, and less saturated fat than sedentary participants (P < .01). They also consumed more vitamins (A, C, D, E), beta carotene, and calcium, (P < .01), and ate more fruits and vegetables (P < .001) than sedentary participants. We found these relationships in both New England communities studied. This association between regular physical activity and diet suggests that the relationship of habitual physical activity and chronic disease may be confounded by diet.