Patient choice modules for summaries of clinical effectiveness: a proposal.
Explore the source record for details and available documents.
Biomedical subjects
Publications and source records attributed to D R Rovner.
Explore the source record for details and available documents.
Patient decision support (PDS) tools or decision aids have been developed as adjuncts to the clinical encounter. Their aim is to support evidence-based patient choice. Clinical trials of PDS tools have used an array of outcome measures to determine efficacy, including knowledge, satisfaction, health status and consistency between patient choice and values. This paper proposes that the correlation between 'subjective expected utility' (SEU) and decision may be the best primary endpoint for trials. SEU is a measure usually used in behavioural decision theory. The paper first describes how decision support tools may use decision analysis to structure the presentation of evidence and guide patient decision-making. Uses of expected utility (EU) are suggested for evaluating PDS tools when improving population health status is the objective. SEU is the theoretically better measure when internal consistency of patient choices is the objective.
BACKGROUND: Patient Decision Support (PDS) tools assist patients in using medical evidence to make choices consistent that are with their values and in using evidence about consequences of medical alternatives. OBJECTIVE: To evaluate a PDS intervention for perimenopausal hormone replacement therapy. We assessed the impact of the PDS on (1) consistency between the decision to take estrogen replacement therapy (ERT) or progesterone/estrogen replacement therapy (PERT) and the expected utility of treatment and (2) likelihood to take ERT and PERT pre- and postintervention. DESIGN: Content of the PDS was standardized. Randomized trial of three intensities of intervention: (1) brochure; (2) lecture/discussion; and (3) active decision support. SUBJECTS: Participants were perimenopausal community volunteers between the ages of 40 and 65 (n = 248). MEASURES: (1) Consistent with values (correlation between expected utility (EU) and likelihood of taking hormones); and (2) Likelihood to take hormone replacement therapy. RESULTS: (1) The brochure group was less consistent with the decision analytic model than the lecture/discussion and active decision support groups. (2) Influence on decisions: PDS tools increased the number of women certain about whether or not to take hormones. There were no differences among experimental groups. Of 99 women uncertain about ERT pre-PDS, 65% changed. Twenty-one (32%) decided against ERT and 44 (68%) decided for ERT. (3) More intensive interventions produced modest gains in a normative direction. CONCLUSIONS: PDSs using any of 3 formats reduce uncertainty and assist women to make informed decisions. Increased consistency with decision analytic models appears to be driven by better estimates of likelihood of outcomes.
It is generally believed that bone mineral density (BMD) is increased in obese subjects, but the effect of weight loss on BMD has not been well studied. Therefore, we evaluated BMD among 11 obese women (mean age 45.5 +/- 14.2 years) before and after weight loss achieved by ingesting an 800 calorie diet for 12 weeks. BMD measurements were made at baseline, 6 months and 1 year intervals. Urinary hydroxyproline:creatinine (H:Cr), calcium:creatinine (Ca:Cr) ratios were measured as indices of bone turnover. Mean weight at baseline was 103.8 +/- 15.8 kg and decreased to 83.2 +/- 12.2 at six months and was 85.8 +/- 14.2 kg at one year. Total body, hip and lumbar spine BMD were 1.12 +/- 0.07, .87 +/- 0.11, and 1.02 +/- 0.12 gm/cm2, respectively. Total body BMD was significantly lower at 12 months compared to baseline. No significant change was observed in BMD of the lumbar spine. There was also a significant decrease in hip BMD at six months and 12 months compared to baseline. H:Cr and Ca:Cr ratios did not change over time. We conclude that weight loss achieved by VLCD is accompanied by a statistically significant change in BMD, but the BMD remained in the normal range.
Evaluation of thyroid nodules remains a challenge for primary care physicians. To include or exclude the presence of malignancy in a thyroid nodule, radioisotope scan, ultrasound, and fine-needle aspiration biopsy of the thyroid generally are used. The objectives of this study were to determine the utility and cost effectiveness of fine-needle aspiration biopsy of solitary thyroid nodules in a community setting; to compare the cost of fine-needle aspiration biopsy with that of radioisotope scan and ultrasound; and to determine whether the practice of obtaining radioisotope scans and ultrasound has changed in the 1990s compared with the 1980s. Patients were referred by community physicians to university-based endocrinologists for evaluation of thyroid nodules. Many of the patients had previously undergone radioisotope scans and ultrasound scans at the discretion of their primary care physicians. All patients underwent fine-needle aspiration biopsy. The biopsy results were evaluated prospectively, and the practice of community physicians' obtaining radioisotope scans and ultrasound scans was compared for the 1980s and 1990s. Eighty-three patients underwent 104 biopsies. In 20 biopsies the specimens were inadequate; the others showed 70 benign, 9 suspicious, and 4 malignant lesions. All four patients with biopsy findings read as malignant were found to have malignant growth at surgical procedures. Two benign biopsy findings were false-negative results. Malignant growth was correctly diagnosed later for one patient at a second biopsy and for the other because of growth of the nodule. The cost of 104 biopsies was $20,800. The cost of radioisotope scans was $22,400, and the cost of ultrasound scans was $10,640. The frequency of obtaining radioisotope scans (84.5% vs 77%) and ultrasound scans (65% vs 45%) was slightly higher in the 1990s compared with the 1980s. Fine-needle aspiration biopsy is a safe and cost effective initial evaluation modality for smaller community-based centers, as it is at large tertiary centers. The cost incurred ($33,040) in obtaining the radioisotope scans and ultrasound scans could have been saved if fine-needle aspiration biopsy had been used as the initial diagnostic procedure for evaluation of these nodules. Although radioisotope scan and ultrasound scan are of little diagnostic help in the evaluation of thyroid nodules, they continued to be obtained at a high frequency during the last decade.
We evaluated serum concentrations of apoprotein (APO) A1, B, total cholesterol, triglycerides, high density cholesterol (HDL-C), and low density cholesterol (LDL-C) in twelve obese subjects whose body mass index (BMI) was > or = 30 before and after a clinically significant weight loss was obtained utilizing a very-low calorie diet (VLCD) consisting of liquid protein (Optifast) providing 800 calories a day. At baseline, the mean weight +/- SD was 119.77 Kg and decreased significantly to 89.29 +/- 13.46 Kg by 24 weeks. Statistically significant reductions of APO-A1, APO-B, total cholesterol, and triglyceride concentrations were also observed along with the weight loss. LDL-C decreased from 156.0 +/- 55.9 mg/dL to 122.5 +/- 42.2 mg/dL (4.03 +/- 1.4 to 3.16 +/- 1.1 mmol/L), but this difference was not statistically significant. There was no significant change in the HDL-C and the ratios of APO-A1 to APO-B. We conclude that the use of VLCD is associated with changes in the lipid pattern that lower the cardiovascular risk profile in addition to the beneficial effects of weight loss itself.
OBJECTIVE: To evaluate the clinical utility of Micral strips for detection of microalbuminuria. DESIGN: One hundred three urine samples were tested by Micral strips for the presence of microalbuminuria, and the results were compared with the immunonephelometric method. SETTING: Endocrine diabetes clinic in a university-affiliated outpatient facility and the associated clinical laboratory. PATIENTS: Sixty-seven, 24-hour urine samples were obtained from 64 patients with diabetes. Thirty-six urine samples were obtained from normal controls; 22 of these were 24-hour samples and 14 were overnight samples. MAIN OUTCOME MEASURE: Concordance of results obtained by the two methods for the presence or absence of microalbuminuria. RESULTS: All 36 control subjects and 44 urine samples from diabetic patients had normal albumin excretion (<15 mg/24 h) by the immunonephelometric method. Seventy-eight of these were correctly identified as negative by Micral strips, giving a specificity of 97.5%. There were 23 samples with albumin excretion of more than 16 mg/24 h. Sixteen of these were correctly identified, giving a sensitivity of 69.5%. There were 16 samples with albumin excretion of 30 mg/24 h or more; 14 of these were correctly identified by Micral strips, and two were false negatives, giving a sensitivity of 87.7%. However, when urine samples with albumin concentrations of less than 11 mg/L were excluded, the Micral strips correctly read 21 out of 23 samples, giving a sensitivity of 91.3%. CONCLUSIONS: The specificity of Micral strips for detection of albuminuria in 24-hour urine samples is high (97.5%), but the sensitivity is low, ranging from 69.5% to 87.7%. The sensitivity was greatly improved when urine samples with albumin concentrations of less than 11 mg/L were excluded.
OBJECTIVES: To determine the prevalence of QT interval prolongation in patients referred to an outpatient clinic for treatment of obesity; and to describe the change in the QT interval during rapid weight loss with a very-low-calorie diet. DESIGN: Retrospective and prospective review of charts and electrocardiograms. SUBJECTS: Five hundred twenty-two obese patients (411 female, 112 males) with a mean age 44 (18-78 y) and a mean initial weight of 116 kg (63-285 kg) completing 26 weeks of treatment between September, 1989 through to December, 1993. MEASUREMENTS: We reviewed the EKGs of all patients and serially monitored the QTc if greater than 0.44 s or if more than 23 kg was lost during treatment. The QTc interval was calculated with Bazzett's formula using both a manual method and an automated software program. In some patients, body composition was measured by hydrodensitometry. RESULTS: The QTc interval before treatment was 0.42 +/- 0.026 s by manual measurement and 0.41 +/- 0.021 s by automated measurement. Forty-one to 53% of patients showed a QTc interval of greater than 0.42 s and 10-24% demonstrated moderate prolongation (> 0.44 s). In those patients for whom repeat EKG were performed, QTc showed shortening with weight loss by both methods (mean +/- s.e. of 0.42 +/- 0.003 to 0.41 +/- 0.003 s, P < 0.01 manually and 0.41 +/- 0.003 to 0.40 +/- 0.003 s, p < 0.005 by automated program). Analyses were repeated excluding 179 patients with a cardiovascular-related diagnosis or intraventricular block and the results were similar. By regression analysis, gender and fat mass (FM) percentage above normal predicted the QTc. CONCLUSIONS: QT Interval prolongation is common in obesity. For each 50% increase in FM% above normal, there is a 5 ms increase in the QTc above a 'normal' upper limit of 0.40 and 0.38 s in women and men, respectively. Moreover, the QT interval shortens with weight loss. This change may represent an additional benefit of weight loss along with the improvement in other cardiovascular risk factors.
Diabetic nephropathy affects approximately 35% of patients with insulin-dependent diabetes mellitus and about 5% to 10% of patients with non-insulin-dependent diabetes mellitus. Diabetic kidney disease can be detected early by special methods that measure microalbuminuria in the range of 15 to 140 micrograms/min or 0.02 to 0.20 g/d. Specific strategies have been proved beneficial in patients with this early diabetic kidney disease, also called "incipient nephropathy." Protein restriction, tight control of blood sugar and blood pressure, and specific therapy with angiotensin-converting enzyme inhibitors have been shown independently to reduce microalbuminuria and preserve renal function in these patients. We review the pathogenesis of diabetic kidney disease and discuss the effects of these treatment strategies on the renal, cardiovascular, metabolic, and other abnormalities found in early (incipient) and overt diabetic nephropathy.
Aminoglutethimide (Ag) is a potent aromatase-enzyme inhibitor used in the treatment of patients with breast cancer. In the past, it has been administered in doses of 1,000 mg/d (usually with 40 mg hydrocortisone). At these dose levels, the drug also affects multiple cytochrome P-450 enzymes, including enzymes for adrenal steroid biosynthesis. Recently, lower-dose regimens (500 mg/d) of Ag have been found to be just as effective for breast cancer therapy, but less toxic than the higher conventional dose. There is limited information on the adrenal effects at the lower dosages, and it is not known whether these effects are clinically significant. We measured basal and synthetic corticotropin (Cortrosyn)-stimulated levels of adrenal steroids in postmenopausal breast cancer patients before and during treatment with low-dose Ag (500 mg/d) administered without a glucocorticoid preparation. Basal levels of progesterone, 17-OH progesterone, and 11-deoxycortisol were higher after 2 months' treatment (P < .01). After ACTH injection, peak levels of progesterone and 17-OH progesterone were higher (P < .01), but in contrast, peak levels of 18-OH corticosterone were lower during treatment (P < .02). Basal and peak levels of cortisol, aldosterone, and all other adrenal steroids were unchanged during treatment. We conclude that low-dose Ag treatment leads to partial inhibition of the 21-hydroxylase, 11-hydroxylase, and 18-hydroxylase adrenal enzymes. Since cortisol and aldosterone secretion remained normal with minimal shunting to or accumulation of adrenal androgen compounds, we believe that the mild inhibition was compensated for by further endogenous ACTH stimulation.
Glucose tolerance deteriorates with advancing age. According to the currently accepted National Diabetes Data Group criteria, the prevalence of diabetes increases from 2% in the 20- to 44-year-old age-group to nearly 18% in those aged 65 to 74 years. Although the principles of management of diabetes in the elderly are similar to those applied to young adults, they must be modified according to the coexisting conditions unique to the elderly. Treatment modalities include diet, exercise, oral hypoglycemic agents (sulfonylureas, metformin in the near future), and insulin. A dietary and exercise prescription should be given to all patients, but must be carefully individualized. If nonpharmacologic measures do not succeed, oral hypoglycemic agents or insulin may be used. It should not be assumed that long-term vascular complications of diabetes will not occur in elderly diabetics. An attempt should therefore be made to control hyperglycemia. The Diabetes Control and Complications Trial demonstrated that tight blood glucose control effectively delays the onset and progression of complications in young patients with insulin-dependent diabetes; however, tight control is associated with an increased risk of hypoglycemia. Hypoglycemia in older adults may have serious consequences, such as precipitating a myocardial infarction or stroke. Treatment should thus aim for a reasonable degree of blood glucose control without producing hypoglycemia or seriously interfering with the enjoyment of life. Coexisting renal or hepatic dysfunction or other factors also make the elderly more prone to developing hypoglycemia, which must be carefully avoided. If complications are present, their treatment and efforts to prevent their worsening are important objectives of the therapeutic plan.
"Steroid burst therapy" is commonly used for various acute medical conditions, but its suppressive effect on hypothalamic-pituitary-adrenocortical (HPA) function and the time period for recovery of HPA function is not fully known. We therefore evaluated the HPA function in 10 normal adults before and after a short burst of Prednisone (40 mg/three times daily for 3 days, then tapered over the next 4 days). HPA function was evaluated by iv administration of 100 micrograms of ovine CRH (oCRH) and blood samples for ACTH and cortisol assay were obtained at -30,0,10,15,30,60,90, and 120 min. On another day, 250 micrograms synthetic ACTH (Cosyntropin) were given iv and blood samples for cortisol were obtained at 0,30,60, and 90 min. Basal and peak levels of ACTH and cortisol before and 1,2, and 3 weeks after discontinuation of prednisone in response to oCRH iv are shown below (see Table 1). All values are mean (SEM). Peak levels of cortisol after iv administration of Cosyntropin at week 0 were 922(56.8), week 1 899(63.7), week 2 861(70.9), and week 3 855(53.0). There was no significant difference noted in the levels of ACTH and cortisol in response to oCRH before and after prednisone treatment. Pre- and posttreatment responses of cortisol to Cosyntropin administration were also similar. In addition, cumulative responses (area under the curve) and the change from baseline (delta) before and after administration of prednisone were similar for ACTH and cortisol. We conclude that HPA function is normal 1 week after discontinuation of a short burst of prednisone. These findings suggest that administration of additional steroids may not be required during periods of "stress" for those patients who have previously received similar steroid burst therapy, if at least 1 week has elapsed after such treatment was given.
Spontaneous pregnancy in patients with premature ovarian failure, although previously reported, remains quite rare. We report a patient with premature ovarian failure who spontaneously conceived and carried the pregnancy to full term. A 26 year old, white female with primary hypothyroidism (diagnosed in 1979) was first seen by us in January 1984 for secondary amenorrhea of six months duration. Menarche occurred at age 13 and menstrual periods were regular until May 1983. She experienced hot flashes in the summer of 1983, and also complained of dryness in her vagina. Serum estrogens were low, and serum gonadotropins were elevated. Physical exam was within normal limits except for a slight, diffuse enlargement of the thyroid gland. Antithyroglobulin and anti-microsomal antibodies were positive. She was clinically and bio-chemically euthyroid while taking exogenous L-thyroxine. A diagnosis of premature ovarian failure was, therefore, made. Therapy was initiated with Premarin and Provera, and she began to have regular withdrawal bleeding. In February 1991, she presented to her obstetrician stating that she was pregnant. The course of the pregnancy was uncomplicated. A cesarean section was performed at 40.5 weeks gestation and she delivered a live female infant. This case confirms that, although rare, patients with premature ovarian failure may spontaneously ovulate and conceive. The course of pregnancy in patients with ovarian failure appears to be normal.
A principal method of studying physician practice patterns has been to examine physicians' responses to brief written cases. We have compared this method with practice patterns of the same physicians derived from chart audit. Subjects were 98 family practice residents for whom data were available in actual patient encounters for the workup of asymptomatic hypertension. Short, carefully structured case reports using four cues were designed and a checklist similar to the one used for test ordering in practice was employed. Chart reviews and billing encounter forms were used for comparison. Results indicated residents ordered fewer tests in clinical practice, due, in part, to practice constraints not represented in the written cases. Physicians tend to make the diagnosis of hypertension incrementally in practice, with no one visit adequately representing the point of diagnosis. Studies based on data bases using a patient encounter as the unit of analysis in chronic disease such as hypertension may spuriously underestimate the actual number of tests ordered for the workup. Judgment cases may better reflect the patterns of use of information in a well-defined problem. Prediction of number of tests ordered in the clinical setting has not been established in this case.
The preferred treatment of hyperthyroidism remains controversial. Most of this data is derived from large, university-based medical centers. We report here our experience with treatment of hyperthyroidism in a community setting. This involves 144 patients with hyperthyroidism who were seen over a 10 year period at Michigan State University Clinical Center and were treated in the community hospitals and private physicians' offices, and by community surgeons. Follow-up data were available on 119 of these patients; 105 of them were hyperthyroid because of Graves' disease and multinodular goiter. Patients were encouraged to make their own decisions regarding choice of therapy, as independently as possible. Sixty-five percent of these patients were treated by 131I, 18% by antithyroid drugs, and 17% by surgery. The mean follow-up period was 2.5 years (range 2 months to 19 years). Hyperthyroidism was controlled in 84% of the patients treated by 131I and 83% of the patients treated by surgery. Forty percent of the patients treated by 131I and 33% treated by surgery became hypothyroid. Fifty percent of the patients achieved remission when treated by antithyroid drugs alone. Our results indicate that when patients are encouraged to make their own decisions regarding the treatment of hyperthyroidism, their choices are similar to those of the thyroidologists. Secondly, the results obtained with different modalities of treatment for hyperthyroidism in a community setting are similar to those obtained in university medical centers.
To determine whether a therapeutic dose of iodine 131 affects the results of 24-hour radioactive iodine uptake (RAIU) testing, we reviewed records of hyperthyroid patients previously treated with 131I at Michigan State University and its affiliated hospitals. We identified 26 patients who had had clinical evaluation and determination of the serum thyroxine (T4) level, triiodothyronine resin uptake (T3RU), and RAIU (using 131I) within two weeks before and several months after the therapeutic dose of 131I. Before treatment, all patients had clinical hyperthyroidism, with an elevated T4 level and increased T3RU and RAIU. After treatment with 131I, eight patients (31%) had an RAIU that was discordant with their clinical and biochemical (T4 and T3RU) assessment. In six patients (23%) of the RAIU was inappropriately high, and in two patients (8%) it was inappropriately low. Since we did not identify any other factors known to interfere with the results of RAIU testing, we conclude that a therapeutic dose of 131I, may by itself increase or decrease a subsequent RAIU determination; therefore, after treatment with 131I, RAIU is not a good diagnostic index of thyroid activity.
Several interactions between thymosin (fraction 5) and the endocrine system have been described in animals, but its effects in humans have not been studied. Plasma levels of glucose, insulin and glucagon were obtained in response to a glucose load before and after treatment with thymosin in eight patients. The glucose tolerance was normal in all subjects and remained so following thymosin administration. No significant changes in the levels of glucagon and insulin were obtained. In six patients, baseline levels of plasma cortisol and ACTH were also obtained. Four out of six patients demonstrated an increase in ACTH levels although the mean differences did not achieve statistical significance. Plasma cortisol levels also did not change significantly. We conclude that thymosin given over a short term is unlikely to significantly influence carbohydrate metabolism. Since four out of six patients showed an increase in basal ACTH levels, it suggests that thymosin may affect the hypothalamic pituitary/adrenal axis. However, further study is needed for complete evaluation of these effects.
Elevated human chorionic gonadotropin levels may be an early, and occasionally the only, manifestation of embryonal cell carcinoma. A 24-year-old man presented with gynecomastia, galactorrhea, and elevated beta-human chorionic gonadotropin levels, which led to an extensive (but nonrevealing) search, including computed tomography and selective testicular vein catheterization, for malignancy. Since the testicle was considered as the most likely site of tumor, right orchiectomy and right common iliac lymph node biopsy were performed. The testicle was normal, but the lymph node contained elements of embryonal and choriocarcinoma. Following chemotherapy, beta-human chorionic gonadotropin levels were normalized and the patient appears "cured." This case emphasizes the need for an extensive search for malignancy and consideration of orchiectomy in such instances in order to achieve a favorable outcome.