Factors associated with antibiotic prescribing in a managed care setting: an exploratory investigation

Perceptions of primary health care physicians about the prescription of antibiotics in Saudi Arabia: Based on the model of Theory of planned behaviour

Introduction

Knowledge of different factors that influence physician decision-making in general practice for prescribing antibiotics is vital for successfully implementing antimicrobial stewardship interventions. The Theory of planned behaviour (TPB) is an established model that describes how cognitions drive human behaviour.

Objectives

The present study was performed to identify the socio-cultural and behavioural determinants that affect antibiotics prescription behaviour among primary care physicians and estimate the awareness about antibiotic resistance of public health importance.

Methodology

The questionnaire was divided into three sections: Section-I, which contained demographic content. Section-II contained 21 question items measuring attitudes, subjective norms, perceived behavioural control, and behavioural intentions of the prescribers. Section III had 12 questions related to knowledge about antibiotic resistance. As determined by Cronbach's α for each set of measured constructs of questionnaire scale, the internal consistency was good (α ≥ 0.7). The participants' socio-cultural and behavioural determinants were graded in 3 categories depending upon the mean score, while frequencies were used to estimate antibiotic resistance awareness. The data was analyzed by calculating the one-way ANOVA and by post-hoc Tukey-Kramer HSD multiple comparison test.

Results

Four hundred thirty-four primary health care physicians responded to the questionnaire, and data was statistically investigated. Participants' knowledge of antibiotic-resistant bacteria was poor with greater behavioural intentions to prescribe antibiotics, but a positive attitude and acceptable perceived social pressure towards antibiotics.

Conclusion

Our study showed that awareness towards antibiotic-resistant bacteria among primary care physicians of Saudi Arabia stands insufficient but favourable attitude and satisfactory social pressure towards antibiotics with greater behavioural intentions to prescribe antibiotics. The continued medical education (CME) and frequent training interventions can increase these parameters of rational prescribing practice for antibiotics.

Knowledge, Attitudes and Intentions to Prescribe Antibiotics: A Structural Equation Modeling Study of Primary Care Institutions in Hubei, China

The aim of this paper is to measure the knowledge and attitudes of primary care physicians toward antibiotic prescriptions and their impacts on antibiotic prescribing. A questionnaire survey was conducted on 625 physicians from 67 primary care facilities in Hubei, China. Structural equation modelling (SEM) was applied to test the theoretical framework derived from the Knowledge, Attitudes, and Practices (KAP) theory. Physicians’ knowledge, five sub-types of attitudes, and three sub-types of behavioral intentions towards antibiotic use were measured. Physicians had limited knowledge about antibiotic prescriptions (average 54.55% correct answers to 11 questions). Although they were generally concerned about antibiotic resistance (mean = 1.28, SD = 0.43), and were reluctant to be submissive to pressures from consumer demands for antibiotics (mean = 1.29, SD = 0.65) and the requirements of defensive practice (mean = 1.11, SD = 0.63), there was a lack of motivation to change prescribing practices (mean = −0.29, SD = 0.70) and strong agreement that other stakeholders should take the responsibility (mean = −1.15, SD = 0.45). The SEM results showed that poor knowledge, unawareness of antibiotic resistance, and limited motivation to change contributed to physicians’ high antibiotics prescriptions (p < 0.001). To curb antibiotic over-prescriptions, improving knowledge itself is not enough. The lack of motivation of physicians to change needs to be addressed through a systematic approach.

Determinants of antibiotic prescribing behaviors of primary care physicians in Hubei of China: a structural equation model based on the theory of planned behavior

Background

Over-prescription of antibiotics is prevalent all over the world, contributing to the development of antibiotic resistance. The importance of understanding how physicians prescribe antibiotics is increasingly highlighted for the purpose of promoting good practice. This study aimed to identify factors that shape the antibiotic prescribing behaviors of physicians in primary care based on the theory of planned behavior (TPB).

Methods

Data were collected from 503 prescribers within 65 primary care facilities in Hubei, tapping into four behavioral aspects leading to antibiotic prescribing based on TPB, namely, attitudes (the degree to which a prescriber is in favor of the use of antibiotics), subjective norms (perceived social pressure to which a prescriber is subject in relation to antibiotic prescriptions), perceived control of behaviors (how easy a prescriber feels in making a rational decision on antibiotic prescriptions) and intentions (the degree to which a prescriber is willing to prescribe antibiotics). A total of 440,268 prescriptions were audited to assess physician antibiotic prescribing practices. The four behavioral constructs were further linked with physician's actual use of antibiotics using structural equation modelling (SEM) based on TPB.

Results

On average, 40.54% (SD = 20.82%) of the outpatient encounters resulted in a prescription for an antibiotic given by the participants and 9.81% (SD = 10.18%) of the patients were given two or more antibiotics. The participants showing a more favorable attitude toward antibiotics had a higher intention to prescribe antibiotics (β = 0.226, p < 0.001) and a lower intention to reduce antibiotic use (β = - 0.211, p < 0.001). Those who perceived lower social pressure (β = 0.113, p = 0.030) and higher control over prescribing behaviors (β = 0.113, p = 0.037) reported a higher intention to reduce the use of antibiotics. However, such intention did not translate into prescribing practice (p > 0.05), although stronger perceived behavioral control was directly linked with lower antibiotic prescriptions (β = - 0.110, p = 0.019). Weaker perceived behavioral control was evident in the participants who showed a less favorable attitude toward antibiotics (β = 0.128, p = 0.001).

Conclusion

Antibiotic prescribing practice is not under the volitional control of prescribers in primary care in China. Their rational prescribing practice is likely to be jeopardized by perceived weak control over prescribing behaviors.

Development of a simple 12-item theory-based instrument to assess the impact of continuing professional development on clinical behavioral intentions

BACKGROUND

Decision-makers in organizations providing continuing professional development (CPD) have identified the need for routine assessment of its impact on practice. We sought to develop a theory-based instrument for evaluating the impact of CPD activities on health professionals' clinical behavioral intentions.

METHODS AND FINDINGS

Our multipronged study had four phases. 1) We systematically reviewed the literature for instruments that used socio-cognitive theories to assess healthcare professionals' clinically-oriented behavioral intentions and/or behaviors; we extracted items relating to the theoretical constructs of an integrated model of healthcare professionals' behaviors and removed duplicates. 2) A committee of researchers and CPD decision-makers selected a pool of items relevant to CPD. 3) An international group of experts (n = 70) reached consensus on the most relevant items using electronic Delphi surveys. 4) We created a preliminary instrument with the items found most relevant and assessed its factorial validity, internal consistency and reliability (weighted kappa) over a two-week period among 138 physicians attending a CPD activity. Out of 72 potentially relevant instruments, 47 were analyzed. Of the 1218 items extracted from these, 16% were discarded as improperly phrased and 70% discarded as duplicates. Mapping the remaining items onto the constructs of the integrated model of healthcare professionals' behaviors yielded a minimum of 18 and a maximum of 275 items per construct. The partnership committee retained 61 items covering all seven constructs. Two iterations of the Delphi process produced consensus on a provisional 40-item questionnaire. Exploratory factorial analysis following test-retest resulted in a 12-item questionnaire. Cronbach's coefficients for the constructs varied from 0.77 to 0.85.

CONCLUSION

A 12-item theory-based instrument for assessing the impact of CPD activities on health professionals' clinical behavioral intentions showed adequate validity and reliability. Further studies could assess its responsiveness to behavior change following CPD activities and its capacity to predict health professionals' clinical performance.

Prescription drug request and denial

Purpose

Prescription drug requests and physician denial are important aspects of medical decision making, but little research has been done to identify factors linked to prescription drug request and physician denial. This paper aims to explore factors in relation to patient prescription drug request and provider denial.

Design/methodology/approach

The paper is based on a cross‐sectional study in a nationally representative database of 2,988 individuals. Descriptive and multivariate stepwise conditional logistic regression analyses were conducted.

Findings

Results of multivariate regression models reveal, after adjusting for personal factors, that heart disease, allergy, anxiety, minor chronic conditions, medical seeking behaviors and direct‐to‐consumer advertising (DTCA) were found to be related to prescription drug request. The denied were individuals with arthritis, less prevalent chronic conditions, the uninsured, and African Americans. It was also found that 27.4 percent of the sample requested a prescription drug and about 24 percent of those who segmented for prescriptions were physicians.

Research limitations/implications

DTCA is positively associated with prescription drug requests but the analysis did not support any effect of DTCA on the refusal status. Patients' requests and physician decision making to refuse are somewhat complicated and vary with different medical conditions.

Originality/value

The paper, using nationally representative data, investigates the factors associated with prescription drug request and denial.

Family doctors' views of pharmaceutical sales representatives: assessment scale development

BACKGROUND

The prescribing patterns depend on the physicians' attitudes and their subjective norms towards prescribing a particular drug, as well as on their personal experience with a particular drug. The physicians are affected by their interactions with pharmaceutical industry.

OBJECTIVE

The objectives were to develop a scale for assessment of pharmaceutical sales representatives (PSRs) by the family doctors (FDs) and to determine factors for their evaluation.

METHOD

Cross-sectional anonymous postal study. We included a random sample of 250 Slovenian FDs. Settings. Slovenian FDs' surgeries.

MAIN OUTCOME MEASURE

The score of various items regarding FDs' assessment of PSRs on a 7-point Likert scale.

RESULTS

We got 163 responses (65.2% response rate). The most important characteristic of PSRs, as rated by respondents on the scale from 1 to 7, was the fact that they did not mislead when presenting products' information. The second most important characteristic was the ability to provide objective information about the product. The first three most important characteristics, as rated by the respondents by themselves, were 'Shows good knowledge on the promoted subject', 'Provides objective product information' and 'Makes brief and exact visits'. Cronbach's alpha of the composite scale was 0.844. Factor analysis revealed three PSRs' factors: selling skills, communicating skills and sense of trustworthiness.

CONCLUSION

FDs evaluate PSRs mainly by their managerial skills and trustworthiness. The scale proved to be a reliable tool for assessing PSRs by FDs.

The role of computerized decision support in reducing errors in selecting medicines for prescription: narrative review

This narrative review includes a summary of research examining prescribing errors, prescription decision making and the role computerized decision support plays in this decision-making process. A reduction in medication prescribing errors, specifically a reduction in the selection of inappropriate medications, is expected to result from the implementation of an effective computerized decision support system. Previous research has investigated the impact of the implementation of electronic systems on medication errors more broadly. This review examines the specific characteristics of decision support systems that may contribute to fewer knowledge-based mistakes in prescribing, and critically appraises the large volume of information available on the decision-making process of selecting medicines for prescription. The results highlight a need for work investigating what decision strategies are used by doctors with different levels of expertise in the prescribing of medications. The nature of the relationship between decision support and decision performance is not well understood and future research is needed to determine the mechanisms by which computerized decision support influences medication selection.

Divergent intentions to use antibiotic guidelines: a theory of planned behavior survey

BACKGROUND

To improve physicians' antimicrobial practice, it is important to identify barriers to and facilitators of guideline adherence and assess their relative importance. The theory of planned behavior permits such assessment and has been previously used for evaluating antibiotic use. According to this theory, guideline use is fueled by 3 factors: attitude, subjective norm (perceived social pressure regarding guidelines), and perceived behavioral control (PBC; perceived ability to follow the guideline). The authors aim to explore factors affecting guideline use in their hospital.

METHODS

Starting from their earlier observations, the authors constructed a questionnaire based on the theory of planned behavior, with an additional measure of habit strength. After pilot testing, the survey was distributed among physicians in a major teaching hospital.

RESULTS

Of 393 contacted physicians, 195 completed questionnaires were received (50.5% corrected response rate). Using multivariate analysis, the overall intention toward using antibiotic guidelines was not very predictable (model R (2) = .134). Habit strength (relative weight = .391) and PBC (relative weight = .354) were the principal significant predictors. A moderator effect of respondents' position (staff member v. resident) was found, with staff members' intention being significantly influenced only by habit strength and residents' intention by PBC. Regarding previously identified barriers, education on antibiotics and guidelines was rated unsatisfactory.

CONCLUSIONS

These divergent origins of influence on guideline adherence point to different approaches for improvement. As habits strongly influence staff members, methods that focus on changing habits (e.g., automated decision support systems) are possible interventions. As residents' intention seems to be guided mainly by external influences and experienced control, this may make feedback, convenient guideline formats, and guideline familiarization more suitable.

HEALTH SERVICES RESEARCH METHODS IN PHARMACY: Triangulation

There are many examples of studies in which pharmacy practice researchers have combined different types of data, methods and approaches within a single research project. Although a common practice in research across many disciplines, concerns have been expressed over combining methods underpinned by different philosophical assumptions and conceptual frameworks. The choice of methods made by pharmacy researchers is generally governed by practical considerations of the research and data requirements for fulfilling research objectives rather than these theoretical concerns. This review of the pharmacy practice research literature demonstrates how a wide variety of methods have been combined for different purposes and in a variety of study designs.

Physicians' generic drug prescribing behavior in district hospitals: a case of Phitsanulok, Thailand

Generic prescribing is a sound approach to contain health care costs. However, little is known about physicians’ prescribing patterns in the Thai context.

Objective:

To explore physicians’ generic prescription patterns in district hospitals.

Methods:

Data was collected from three of the eight district hospitals between January and December 2008 (final response rate 37.5%). All participating hospitals were between 30 and 60-bed capacity. The researchers reviewed 10% of total outpatient prescriptions in each hospital.

Results:

A total of 14,500 prescriptions were evaluated. The majority of patients were under universal health coverage (4,367; 30.1%), followed by senior citizens’ health insurance (2,734; 18.9%), and civil servant medical benefit schemes (2,419; 16.7%). Ten thousand six hundred and seventy-one prescriptions (73.6% of total prescriptions) had at least one medication. Among these, each prescription contained 2.85 (SD=1.69) items. The majority of prescriptions (7,886; 73.9%) were prescribed by generic name only. Drugs prescribed by brand names varied in their pharmacological actions. They represented both innovator and branded-generic items. Interestingly, a large number of them were fixed-dose combination drugs. All brand name prescriptions were off patented. In addition, none of the brand-name drugs prescribed were categorized as narrow therapeutic range or any other drug that had been reported to have had problems with generic substitution.

Conclusion:

The majority of prescriptions in this sample were written by generic names. There is room for improvement in brand name prescribing patterns.

Using the theory of reasoned action to determine physicians' intention to measure body mass index in children and adolescents

BACKGROUND

Over the past few decades, childhood obesity has become a major public health issue in the United States. Numerous public and professional organizations recommend that physicians periodically screen for obesity in children and adolescents using the body mass index (BMI). However, studies have shown that physicians infrequently measure BMI in children and adolescents.

OBJECTIVES

The purpose of this study was to use the theory of reasoned action (TRA) to explain physicians' intentions to measure BMI in children and adolescents. The study objectives were to (1) determine if attitude and subjective norm predict physicians' intention to measure BMI in children and adolescents; (2) determine if family physicians and pediatricians differ in terms of theoretical factors; and (3) assess differences in behavioral beliefs, outcome evaluations, normative beliefs, and motivation to comply among physicians based on their level of intention to measure BMI.

METHODS

A cross-sectional mailed survey of 2590 physicians (family physicians and pediatricians) practicing in 4 states was conducted. A self-administered questionnaire was designed that included items related to the TRA constructs. The association between the theoretical constructs was examined using correlation and regression analyses. Student's t test was used to determine differences between family physicians and pediatricians on theoretical constructs and to compare the underlying beliefs of nonintenders with intenders.

RESULTS

The usable response rate was 22.8%. Less than half (44%) of the physicians strongly intended to measure BMI in children and adolescents. Together, the TRA constructs attitude and subjective norm explained up to 49.9% of the variance in intention. Pediatricians had a significantly (P<.01) higher intention to measure BMI as compared to family physicians. There were significant (P<.01) behavioral and normative belief differences between physicians who intend and those who do not intend to measure BMI.

CONCLUSION

The TRA is a useful model in identifying the factors that are associated with physicians' intentions to measure BMI.

Healthcare professionals' intentions and behaviours: a systematic review of studies based on social cognitive theories

BACKGROUND

There is an important gap between the implications of clinical research evidence and the routine clinical practice of healthcare professionals. Because individual decisions are often central to adoption of a clinical-related behaviour, more information about the cognitive mechanisms underlying behaviours is needed to improve behaviour change interventions targeting healthcare professionals. The aim of this study was to systematically review the published scientific literature about factors influencing health professionals' behaviours based on social cognitive theories. These theories refer to theories where individual cognitions/thoughts are viewed as processes intervening between observable stimuli and responses in real world situations.

METHODS

We searched psycINFO, MEDLINE, EMBASE, CIHNAL, Index to theses, PROQUEST dissertations and theses and Current Contents for articles published in English only. We included studies that aimed to predict healthcare professionals' intentions and behaviours with a clear specification of relying on a social cognitive theory. Information on percent of explained variance (R(2)) was used to compute the overall frequency-weighted mean R(2) to evaluate the efficacy of prediction in several contexts and according to different methodological aspects. The cognitive factors most consistently associated with prediction of healthcare professionals' intention and behaviours were documented.

RESULTS

Seventy eight studies met the inclusion criteria. Among these studies, 72 provided information on the determinants of intention and 16 prospective studies provided information on the determinants of behaviour. The theory most often used as reference was the Theory of Reasoned Action (TRA) or its extension the Theory of Planned Behaviour (TPB). An overall frequency-weighted mean R(2) of 0.31 was observed for the prediction of behaviour; 0.59 for the prediction of intention. A number of moderators influenced the efficacy of prediction; frequency-weighted mean R(2) varied from 0.001 to 0.58 for behaviour and 0.19 to 0.81 for intention.

CONCLUSION

Our results suggest that the TPB appears to be an appropriate theory to predict behaviour whereas other theories better capture the dynamic underlying intention. In addition, given the variations in efficacy of prediction, special care should be given to methodological issues, especially to better define the context of behaviour performance.

Sample size for regression analyses of theory of planned behaviour studies: case of prescribing in general practice

OBJECTIVES

Interest has been growing in the use of the theory of planned behaviour (TBP) in health services research. The sample sizes range from less than 50 to more than 750 in published TPB studies without sample size calculations. We estimate the sample size for a multi-stage random survey of prescribing intention and actual prescribing for asthma in British general practice. To our knowledge, this is the first systematic attempt to determine sample size for a TPB survey.

METHODS

We use two different approaches: reported values of regression models' goodness-of-fit (the lambda method) and zero-order correlations (the variance inflation factor or VIF method). Intra-cluster correlation coefficient (ICC) is estimated and a socioeconomic variable is used for stratification. We perform sensitivity analysis to estimate the effects of our decisions on final sample size.

RESULTS

The VIF method is more sensitive to the requirements of a TPB study. Given a correlation of .25 between intention and behaviour, and of .4 between intention and perceived behavioural control, the proposed sample size is 148. We estimate the ICC for asthma prescribing to be around 0.07. If 10 general practitioners were sampled per cluster, the sample size would be 242.

CONCLUSIONS

It is feasible to perform sophisticated sample size calculations for a TPB study. The VIF is the appropriate method. Our approach can be used with adjustments in other settings and for other regression models.

Do self- reported intentions predict clinicians' behaviour: a systematic review

Background

Implementation research is the scientific study of methods to promote the systematic uptake of clinical research findings into routine clinical practice. Several interventions have been shown to be effective in changing health care professionals' behaviour, but heterogeneity within interventions, targeted behaviours, and study settings make generalisation difficult. Therefore, it is necessary to identify the 'active ingredients' in professional behaviour change strategies. Theories of human behaviour that feature an individual's "intention" to do something as the most immediate predictor of their behaviour have proved to be useful in non-clinical populations. As clinical practice is a form of human behaviour such theories may offer a basis for developing a scientific rationale for the choice of intervention to use in the implementation of new practice. The aim of this review was to explore the relationship between intention and behaviour in clinicians and how this compares to the intention-behaviour relationship in studies of non-clinicians.

Methods

We searched: PsycINFO, MEDLINE, EMBASE, CINAHL, Cochrane Central Register of Controlled Trials, Science/Social science citation index, Current contents (social & behavioural med/clinical med), ISI conference proceedings, and Index to Theses. The reference lists of all included papers were checked manually. Studies were eligible for inclusion if they had: examined a clinical behaviour within a clinical context, included measures of both intention and behaviour, measured behaviour after intention, and explored this relationship quantitatively. All titles and abstracts retrieved by electronic searching were screened independently by two reviewers, with disagreements resolved by discussion.

Discussion

Ten studies were found that examined the relationship between intention and clinical behaviours in 1623 health professionals. The proportion of variance in behaviour explained by intention was of a similar magnitude to that found in the literature relating to non-health professionals. This was more consistently the case for studies in which intention-behaviour correspondence was good and behaviour was self-reported. Though firm conclusions are limited by a smaller literature, our findings are consistent with that of the non-health professional literature. This review, viewed in the context of the larger populations of studies, provides encouragement for the contention that there is a predictable relationship between the intentions of a health professional and their subsequent behaviour. However, there remain significant methodological challenges.

Practice nurses' intentions to use clinical guidelines

BACKGROUND

Evidence indicates that, although nurses are increasingly using clinical guidelines to ensure higher quality of care, there is a wide variance in their adherence to them. The utility of the Theory of Planned Behaviour (TPB) has not been previously investigated in explaining this variance in community nursing.

AIM

This paper reports a study whose primary aim was to examine the utility of the TPB in explaining variations in practice nurses' intentions to offer smoking cessation advice in accordance with coronary heart disease guidelines.

METHODS

A cross-sectional survey using a postal questionnaire was carried out. A 52-item questionnaire was administered to 48 practice nurses in England. The questionnaire was designed to assess the components of the TPB, and included measures of intentions to offer smoking cessation advice, self-reported past behaviour, attitudes, subjective norms, perceived behavioural controls (PBCs), behavioural beliefs and evaluations, normative beliefs and motivation to comply, and control beliefs and evaluations.

RESULTS

The TPB explained up to 40% of variance in intentions to offer smoking cessation advice. Attitudes and PBCs were the most important predictors of intention. Among other elements of the TPB, indirect attitudes and indirect PBCs made significant positive contributions to explaining variance in intention.

DISCUSSION

Future trials of interventions to increase practice nurses' adherence to clinical guidelines could attempt to address the elements identified in this study as important factors. Further studies are required to examine the utility of the TPB in predicting practice nurses' behaviour.

Obstetricians' attitudes, subjective norms, perceived controls, and intentions on antibiotic prophylaxis in caesarean section

Over 10% of current births in all countries of the world are delivered by caesarean section. Single-dose ampicillin or cefazolin administered after cord clamping has been proven to be effective for the prevention of post-caesarean infections as indicated in many randomised trials and reviews in the Cochrane Library. This study aimed to determine three determinants of behavioural intention using the theory of planned behaviour: attitudes, subjective norms, and perceived controls. Intentions were examined for five aspects of the use of antibiotic prophylaxis, namely whether or not antibiotics were used, used in all caesarean sections, after rather than before cord clamping, whether ampicillin/cefazolin or broader-spectrum antibiotics were used, and whether single or multiple doses were given. Fifty obstetricians selected from university, regional, and general hospitals in southern Thailand, were surveyed using a questionnaire and in-depth interview. Their intentions to use a single dose and to use in all cases were low, and this was related to negative attitudes and reference groups who did not approve of the single dose. The negative attitude was based on scepticism concerning the applicability of well-equipped trials from the developed world and fear of consequences of post-caesarean infections. Norms carried over from residency training had more long-term influence in their practice than newer information from books or journals. Perceived external controls on their practice were less predictive of intentions. Intentions were only partly predictive of behaviour. Changing attitudes, introducing evidence-based information into residency training and strengthening control systems in the hospital are essential to improve intentions.

Urinary tract infection in women--physician's preferences for treatment and adherence to guidelines: a national drug utilization study in a managed care setting

BACKGROUND

The treatment of urinary tract infection (UTI), the most common bacterial infection in most Western countries, is a global clinical and economic issue. Trimethoprim-sulfamethoxazole (TMP-SMX) and nitrofurantoin are the two drugs currently recommended in clinical guidelines in Israel for uncomplicated UTI in women.

OBJECTIVES

This study evaluates physician preferences for treatment and adherence to guidelines.

METHOD

Data were derived from the electronic records of Leumit Health Fund, one of four health management organizations in Israel. Non-pregnant women aged 18-75 years with a diagnosis of acute cystitis or UTI without risk factors for complicated UTI who were empirically treated with antibiotics from January 2001 to June 2002 were identified. The final sample comprised 7738 physician-patient encounters. Physician prescription behavior was analyzed by evaluating the proportion of treatments with each individual drug. A binary regression model was implemented to identify factors associated with suboptimal adherence to the guidelines.

RESULTS

TMP-SMX was the most frequently prescribed drug (25.81%), followed by nitrofurantoin (14.71%), for a 40.52% rate of adherence to the guidelines [95% confidence interval (CI)=39.42, 41.61]. Drugs from the fluoroquinolone family were prescribed in 22.82% of cases. Prescription behavior was also influenced by non-clinical, non-pharmacological factors, such as physician specialty, geographic setting and patient age.

CONCLUSIONS

The majority of cases of UTI in the present study were not treated according to the current guidelines. Fluoroquinolones, though not recommended and relatively costly, were prescribed extensively. These results highlight the necessity for a remedial education program within the health care system designed to improve adherence to the guidelines for the treatment of UTI in women. As this issue is of global importance, this evaluation may serve as a model for similar studies in other settings or countries.

Predictors of broad-spectrum antibiotic prescribing for acute respiratory tract infections in adult primary care

CONTEXT

Broad-spectrum antibiotics are commonly prescribed, but little is known about the physicians who prescribe and the patients who take these agents.

OBJECTIVE

To identify factors associated with prescribing of broad-spectrum antibiotics by physicians caring for patients with nonpneumonic acute respiratory tract infections (ARTIs).

DESIGN, SETTING, AND PATIENTS

Cross-sectional study using data from the National Ambulatory Medical Care Survey between 1997 and 1999. Information was collected on a national sample of 1981 adults seen by physicians for the common cold and nonspecific upper respiratory tract infections (URTIs) (24%), acute sinusitis (24%), acute bronchitis (23%), otitis media (5%), pharyngitis, laryngitis, and tracheitis (11%), or more than 1 of the above diagnoses (13%).

MAIN OUTCOME MEASURE

Prescription of broad-spectrum antibiotics, defined for this study as quinolones, amoxicillin/clavulanate, second- and third-generation cephalosporins, and azithromycin and clarithromycin.

RESULTS

Antibiotics were prescribed to 63% of patients with an ARTI, ranging from 46% of patients with the common cold or nonspecific URTIs to 69% of patients with acute sinusitis. Broad-spectrum agents were chosen in 54% of patients prescribed an antibiotic, including 51% of patients with the common cold and nonspecific URTIs, 53% with acute sinusitis, 62% with acute bronchitis, and 65% with otitis media. Multivariable analysis identified several clinical and nonclinical factors associated with choice of a broad-spectrum agent. After adjusting for diagnosis and chronic comorbid illnesses, the strongest independent predictors of broad-spectrum antibiotic prescribing were physician specialty (odds ratio [OR], 2.4; 95% confidence interval [CI], 1.6-3.5 for internal medicine physicians compared with general and family physicians) and geographic region (OR, 2.6; 95% CI, 1.4-4.8 for Northeast and OR, 2.4; 95% CI, 1.4-4.2 for South [both compared with West]). Other independent predictors of choosing a broad-spectrum agent included black race, lack of health insurance, and health maintenance organization membership, each of which was associated with lower rates of broad-spectrum prescribing. Patient age, sex, and urban vs rural location were not significantly associated with prescribing choice.

CONCLUSIONS

Broad-spectrum antibiotics are commonly prescribed for the treatment of ARTIs, especially by internists and physicians in the Northeast and South. These high rates of prescribing, wide variations in practice patterns, and the strong association of nonclinical factors with antibiotic choice suggest opportunities to improve prescribing patterns.