A qualitative approach to primary care research: the long interview

Qualitative research methods offer useful complements to the traditional epidemiologic approaches most commonly used by primary care researchers. Narrative texts generated through open-ended interviews are one form of qualitative data that may provide useful insights into primary care issues that are otherwise overlooked by more structured designs. This paper presents a step-by-step process for conducting a long interview, a qualitative approach with distinct potential for primary care research. Advantages and limitations of the long interview and its relationship to other qualitative methods are also discussed.

Hospitalization experience of Navajo subjects with type II diabetes and matched controls: an historical cohort study

Using an historical cohort study design with a 12 year follow-up, we found that 77 Navajo adults with type II diabetes mellitus were hospitalized at a rate of 335 hospitalizations per 1000 patient years compared to a rate of 167 hospitalizations per 1000 patient years for 77 age, sex, and residence matched non-diabetic controls, yielding a risk ratio of 2.0. Using matched pairs analysis (sign test), the observed difference in number of hospital admissions is statistically significant (z = 2.30, p less than 0.05). The average duration of hospitalization, however, was not statistically different in matched pairs analysis (z = 0.95, p greater than 0.05). The 136 excess hospitalizations of the diabetic subjects included 45 admissions for poor metabolic control of diabetes, 50 excess admissions for infectious disease, and 26 excess admissions for conditions of the heart, eye, kidney, or non-traumatic amputation. In multivariate analyses, variables found to be associated with greater hospitalization experience among the 77 diabetic subjects in the 12 years follow-up period included older age at entry to the study, poorer metabolic control early in the study period, and presence of diabetic complications.

The individual over time: time series applications in health care research

This paper presents a summary and a brief theoretical introduction to time series ARIMA modeling of single subject data. Time series, a statistical technique that may be appropriate when data are measured repeatedly and at nearly equal intervals of time, has potential research applications in the study of chronic diseases such as diabetes, hypertension, and herpes simplex. Both intervention models and multivariate models are covered, with examples illustrating the utility of time series techniques in chronic disease research. Time series modeling of a subject with diabetes before and after being placed on a regimen of chlorpropamide is used to demonstrate the potential of intervention analysis. Multivariate time series techniques are illustrated by modeling the relationship between exercise and blood glucose, and by modelling the relationship between psychosocial distress and lymphocyte subsets of the cellular immune system.

Effects of age, education, and physician advice on utilization of screening mammography

We investigated the utilization of mammography as a screening test for breast cancer in a middle-income Connecticut suburban community of 30,000 people. The sampling frame was community-dwelling women aged 30 years and over who had telephones. Random digit telephone survey methods were used to identify a sample of 470 eligible subjects. Of those eligible to be included, 350 or 74.4% completed the interview. Analysis of data from the 171 respondents aged 50 years or greater indicated that women aged 65-80 years had a significantly lower rate of screening mammography than did women aged 50-64 years (means 2 = 6.6, P = .01). When further analysis was done to take into account the effects of education and of income on these rates, the association of age with mammography utilization was no longer statistically significant. Among women who recalled their physician advising a mammogram, 88% had had one performed. Among women who could not recall their physician advising a mammogram, 7% had had one. The impact of physician advice was statistically significant (means 2 = 110.3, P less than .001). Physicians recommended screening mammography less for patients with low level of education (means 2 = 21.6, P less than .001), low income (X2 = 7.8, df = 2, P = .02) and greater age (means 2 = 14.2, P = .003). We conclude that utilization of screening mammography in the community studied is related more strongly to education and to income than to age. The bivariate association of mammography utilization with age may be attributable to a cohort effect, rather than an age effect.(ABSTRACT TRUNCATED AT 250 WORDS)

Comparison of metabolic control among diabetic subjects at two clinics

A cross-sectional study was conducted to investigate whether glucose control in 20 non-insulin-dependent diabetic subjects seen at a remote rural clinic was comparable to control achieved in 66 diabetic subjects seen at the regional hospital clinic. Sampling was done to assure that study subjects were representative of all care-seeking, diagnosed diabetics in a well-defined Navajo community. The two groups of patients were comparable in terms of age, sex, and duration of diabetes from time of diagnosis. Compliance with care, hospitalization rates, and complication rates were similar in each group. Results showed no significant differences in glycemic control between the rural clinic (mean fasting plasma glucose = 177, mean random plasma glucose = 227) and the regional hospital clinic (mean fasting plasma glucose = 187, mean random plasma glucose = 249). The percentages of diabetics under "acceptable" control by American Diabetes Association guidelines was 40 percent at the rural clinic and 29 percent at the hospital clinic (P greater than .05). The authors conclude that adequacy of glycemic control in diabetics is not compromised by providing care at a remote rural clinic.

Metabolic control in non-insulin-dependent diabetes mellitus: factors associated with patient outcomes

We conducted a cross-sectional study to assess the association of various demographic and medical-care variables with metabolic outcomes in non-insulin-dependent diabetic subjects. The study population was representative of the diagnosed care-seeking diabetic population of a defined geographic community on the Navajo reservation in Arizona. The dependent variable metabolic control was measured as the mean of all random plasma glucose values obtained only at scheduled diabetes clinic visits over 2 yr. Multivariate analysis of the data showed that better metabolic control was most strongly associated with compliance with scheduled appointments. Mode of treatment was also associated with metabolic control. Other variables tested, including source of care, age, sex, duration of diabetes, presence of complications, and weight change, were not associated with metabolic control. The strongest analysis of covariance model with demographic and medical-care variables accounted for 39% of the variance in metabolic control. The analysis suggests that other variables, possibly including several psychosocial variables, need to be assessed for their contribution to metabolic control.