Michael K. Park, MD

Evidence summary

An RCT demonstrated increased prescribing rates when unannounced, standardized patients who imitated symptoms of either major depression or adjustment disorder requested a prescription. In 298 visits, 152 family physicians and internists prescribed antidepressants at significantly different rates when patients mimicking major depression requested brand-specific (53%), general (76%), or no medication (31%) (P<.001). Corresponding rates for adjustment disorder were 55%, 39%, and 10% (P<.001). The study found no difference in prescribing rates between family physicians and internists or between male and female physicians.

For patients presenting with depression, physician fidelity to minimum recommended treatment (defined as any combination of antidepressants, mental health referral, and 2-week follow-up) was 90% for patients making a brand-specific request, 98% for those making a general request, and 56% for those making no request (P<.001).¹

Patients who ask for DTCA meds are more likely to get a prescription
A cross-sectional survey compared prescribing decisions by 38 US and 40 Canadian physicians for 1431 patients. Most physicians fulfilled requests for DTCA medications (US 78%, Canadian, 72%). Patients who requested a DTCA medication were far more likely to receive a new prescription (DTCA or other) than patients who didn’t (odds ratio [OR]=16.9; 95% confidence interval [CI], 7.5-38.2). Although DTCA is illegal in Canada, market contamination seems likely, because the study was done in Vancouver, British Columbia.

US patients made more requests for DTCA medications (OR=2.2; 95% CI, 1.2-4.1) than Canadian patients. When patients made DTCA-related requests, physicians considered 50% of new prescriptions to be only “possible” or “unlikely” choices for other similar patients, compared with only 12.4% when patients didn’t make such requests (P<.001).²

DTCA drugs often trump other options
A national telephone survey of 3000 adults found that 35% of respondents were prompted by DTCA to discuss the medication or related health concern at a doctor visit. Of these, 72.9% reported receiving a new prescription, and 43% of them were for the advertised drug.³

Another survey of 643 physicians showed that 39% of visits influenced by DTCA resulted in new prescriptions for the advertised drug. Physician stated they prescribed the DTCA medication because they wanted to:

• prescribe the most effective drug (46%),
• accommodate the request despite other equally effective options (48%), or
• accommodate the request despite more effective options (5.5%).

Medications other than the advertised drug were prescribed during 22% of DTCA influenced visits.⁴

DTCA and prescriptions: Mixed results
A time-series analysis examined the relationship between advertising expenditures for several types of drugs and prescriptions written from 1992 to 1997.⁵

Advertising for antilipemic drugs was positively associated with prescriptions for both antilipemics in general (41 prescriptions for every $1000 of advertising; P=.003) and Zocor in particular (23/$1000; P<.001). Advertising for antihistamines in general and Claritin in particular were both positively associated with prescriptions for Claritin (general advertising: 24/$1000; P=.004; Claritin-specific advertising: 45/$1000; P=.005).

Advertising for acid-peptic disorder medications was inversely associated with Zantac prescriptions (-59 prescriptions/$1000, P=.004). This finding may be related to the emergence of, and advertising for, proton pump inhibitors during this time.

No relationship between advertising and prescribing was found for antihypertensives or medications for benign prostatic hypertrophy.

More frequent advertising doesn’t necessarily mean more prescriptions
Another time-series analysis examined the relationship between the frequency of local DTCA for cyclooxygenase-2 (COX-2) inhibitors and prescriptions for COX-2 inhibitors in the corresponding months. DTCA was not significantly associated with Celebrex prescriptions, but was slightly associated with Vioxx prescriptions (P=.04, 10-fold increase in DTCA associated with a 0.5% increase in prescriptions; P=.04). Practices farther than 100 miles from a media market and those that prescribed Vioxx infrequently were excluded.⁶ (In 2004, Vioxx was withdrawn from the US and worldwide markets.)

But patient questions about a specific ad get results
A survey of 2929 patients regarding the appropriateness of COX-2 inhibitors showed that 78% of patients who asked their physician about a COX-2 advertisement received a prescription for a COX-2 inhibitor (instead of a nonsteroidal anti-inflammatory drug) compared with 43% of all other patients.⁷

Recommendations

The American Medical Association (AMA) encourages physicians approached by patients about advertised medications to initiate a dialogue in order to enhance the patient’s understanding of the underlying condition. Request ed medications should be prescribed only if indicated and cost effective relative to other options.

The AMA also recommends that physicians report (to the pharmaceutical company or the US Food and Drug Administration) ads that are inaccurate, incomplete, or imbalanced, and ads that don’t enhance patient education or encourage patients to have a discussion with their physician.⁸

Evidence summary

An RCT demonstrated increased prescribing rates when unannounced, standardized patients who imitated symptoms of either major depression or adjustment disorder requested a prescription. In 298 visits, 152 family physicians and internists prescribed antidepressants at significantly different rates when patients mimicking major depression requested brand-specific (53%), general (76%), or no medication (31%) (P<.001). Corresponding rates for adjustment disorder were 55%, 39%, and 10% (P<.001). The study found no difference in prescribing rates between family physicians and internists or between male and female physicians.

For patients presenting with depression, physician fidelity to minimum recommended treatment (defined as any combination of antidepressants, mental health referral, and 2-week follow-up) was 90% for patients making a brand-specific request, 98% for those making a general request, and 56% for those making no request (P<.001).¹

Patients who ask for DTCA meds are more likely to get a prescription
A cross-sectional survey compared prescribing decisions by 38 US and 40 Canadian physicians for 1431 patients. Most physicians fulfilled requests for DTCA medications (US 78%, Canadian, 72%). Patients who requested a DTCA medication were far more likely to receive a new prescription (DTCA or other) than patients who didn’t (odds ratio [OR]=16.9; 95% confidence interval [CI], 7.5-38.2). Although DTCA is illegal in Canada, market contamination seems likely, because the study was done in Vancouver, British Columbia.

US patients made more requests for DTCA medications (OR=2.2; 95% CI, 1.2-4.1) than Canadian patients. When patients made DTCA-related requests, physicians considered 50% of new prescriptions to be only “possible” or “unlikely” choices for other similar patients, compared with only 12.4% when patients didn’t make such requests (P<.001).²

DTCA drugs often trump other options
A national telephone survey of 3000 adults found that 35% of respondents were prompted by DTCA to discuss the medication or related health concern at a doctor visit. Of these, 72.9% reported receiving a new prescription, and 43% of them were for the advertised drug.³

Another survey of 643 physicians showed that 39% of visits influenced by DTCA resulted in new prescriptions for the advertised drug. Physician stated they prescribed the DTCA medication because they wanted to:

• prescribe the most effective drug (46%),
• accommodate the request despite other equally effective options (48%), or
• accommodate the request despite more effective options (5.5%).

Medications other than the advertised drug were prescribed during 22% of DTCA influenced visits.⁴

DTCA and prescriptions: Mixed results
A time-series analysis examined the relationship between advertising expenditures for several types of drugs and prescriptions written from 1992 to 1997.⁵

Advertising for antilipemic drugs was positively associated with prescriptions for both antilipemics in general (41 prescriptions for every $1000 of advertising; P=.003) and Zocor in particular (23/$1000; P<.001). Advertising for antihistamines in general and Claritin in particular were both positively associated with prescriptions for Claritin (general advertising: 24/$1000; P=.004; Claritin-specific advertising: 45/$1000; P=.005).

Advertising for acid-peptic disorder medications was inversely associated with Zantac prescriptions (-59 prescriptions/$1000, P=.004). This finding may be related to the emergence of, and advertising for, proton pump inhibitors during this time.

No relationship between advertising and prescribing was found for antihypertensives or medications for benign prostatic hypertrophy.

More frequent advertising doesn’t necessarily mean more prescriptions
Another time-series analysis examined the relationship between the frequency of local DTCA for cyclooxygenase-2 (COX-2) inhibitors and prescriptions for COX-2 inhibitors in the corresponding months. DTCA was not significantly associated with Celebrex prescriptions, but was slightly associated with Vioxx prescriptions (P=.04, 10-fold increase in DTCA associated with a 0.5% increase in prescriptions; P=.04). Practices farther than 100 miles from a media market and those that prescribed Vioxx infrequently were excluded.⁶ (In 2004, Vioxx was withdrawn from the US and worldwide markets.)

But patient questions about a specific ad get results
A survey of 2929 patients regarding the appropriateness of COX-2 inhibitors showed that 78% of patients who asked their physician about a COX-2 advertisement received a prescription for a COX-2 inhibitor (instead of a nonsteroidal anti-inflammatory drug) compared with 43% of all other patients.⁷

Recommendations

The American Medical Association (AMA) encourages physicians approached by patients about advertised medications to initiate a dialogue in order to enhance the patient’s understanding of the underlying condition. Request ed medications should be prescribed only if indicated and cost effective relative to other options.

The AMA also recommends that physicians report (to the pharmaceutical company or the US Food and Drug Administration) ads that are inaccurate, incomplete, or imbalanced, and ads that don’t enhance patient education or encourage patients to have a discussion with their physician.⁸

Evidence summary

An RCT demonstrated increased prescribing rates when unannounced, standardized patients who imitated symptoms of either major depression or adjustment disorder requested a prescription. In 298 visits, 152 family physicians and internists prescribed antidepressants at significantly different rates when patients mimicking major depression requested brand-specific (53%), general (76%), or no medication (31%) (P<.001). Corresponding rates for adjustment disorder were 55%, 39%, and 10% (P<.001). The study found no difference in prescribing rates between family physicians and internists or between male and female physicians.

For patients presenting with depression, physician fidelity to minimum recommended treatment (defined as any combination of antidepressants, mental health referral, and 2-week follow-up) was 90% for patients making a brand-specific request, 98% for those making a general request, and 56% for those making no request (P<.001).¹

Patients who ask for DTCA meds are more likely to get a prescription
A cross-sectional survey compared prescribing decisions by 38 US and 40 Canadian physicians for 1431 patients. Most physicians fulfilled requests for DTCA medications (US 78%, Canadian, 72%). Patients who requested a DTCA medication were far more likely to receive a new prescription (DTCA or other) than patients who didn’t (odds ratio [OR]=16.9; 95% confidence interval [CI], 7.5-38.2). Although DTCA is illegal in Canada, market contamination seems likely, because the study was done in Vancouver, British Columbia.

US patients made more requests for DTCA medications (OR=2.2; 95% CI, 1.2-4.1) than Canadian patients. When patients made DTCA-related requests, physicians considered 50% of new prescriptions to be only “possible” or “unlikely” choices for other similar patients, compared with only 12.4% when patients didn’t make such requests (P<.001).²

DTCA drugs often trump other options
A national telephone survey of 3000 adults found that 35% of respondents were prompted by DTCA to discuss the medication or related health concern at a doctor visit. Of these, 72.9% reported receiving a new prescription, and 43% of them were for the advertised drug.³

Another survey of 643 physicians showed that 39% of visits influenced by DTCA resulted in new prescriptions for the advertised drug. Physician stated they prescribed the DTCA medication because they wanted to:

• prescribe the most effective drug (46%),
• accommodate the request despite other equally effective options (48%), or
• accommodate the request despite more effective options (5.5%).

Medications other than the advertised drug were prescribed during 22% of DTCA influenced visits.⁴

DTCA and prescriptions: Mixed results
A time-series analysis examined the relationship between advertising expenditures for several types of drugs and prescriptions written from 1992 to 1997.⁵

Advertising for antilipemic drugs was positively associated with prescriptions for both antilipemics in general (41 prescriptions for every $1000 of advertising; P=.003) and Zocor in particular (23/$1000; P<.001). Advertising for antihistamines in general and Claritin in particular were both positively associated with prescriptions for Claritin (general advertising: 24/$1000; P=.004; Claritin-specific advertising: 45/$1000; P=.005).

Advertising for acid-peptic disorder medications was inversely associated with Zantac prescriptions (-59 prescriptions/$1000, P=.004). This finding may be related to the emergence of, and advertising for, proton pump inhibitors during this time.

No relationship between advertising and prescribing was found for antihypertensives or medications for benign prostatic hypertrophy.

More frequent advertising doesn’t necessarily mean more prescriptions
Another time-series analysis examined the relationship between the frequency of local DTCA for cyclooxygenase-2 (COX-2) inhibitors and prescriptions for COX-2 inhibitors in the corresponding months. DTCA was not significantly associated with Celebrex prescriptions, but was slightly associated with Vioxx prescriptions (P=.04, 10-fold increase in DTCA associated with a 0.5% increase in prescriptions; P=.04). Practices farther than 100 miles from a media market and those that prescribed Vioxx infrequently were excluded.⁶ (In 2004, Vioxx was withdrawn from the US and worldwide markets.)

But patient questions about a specific ad get results
A survey of 2929 patients regarding the appropriateness of COX-2 inhibitors showed that 78% of patients who asked their physician about a COX-2 advertisement received a prescription for a COX-2 inhibitor (instead of a nonsteroidal anti-inflammatory drug) compared with 43% of all other patients.⁷

Recommendations

The American Medical Association (AMA) encourages physicians approached by patients about advertised medications to initiate a dialogue in order to enhance the patient’s understanding of the underlying condition. Request ed medications should be prescribed only if indicated and cost effective relative to other options.

The AMA also recommends that physicians report (to the pharmaceutical company or the US Food and Drug Administration) ads that are inaccurate, incomplete, or imbalanced, and ads that don’t enhance patient education or encourage patients to have a discussion with their physician.⁸

Evidence summary

A systematic review conducted for the US Preventive Services Task Force (USPSTF) addressed the use of aspirin for primary prevention of colorectal carcinomas (CRC) and colorectal adenomas (CRA).

Pooled data from 2 randomized-controlled trials (RCTs) with a total of 61,947 patients showed no decrease in CRC incidence (relative risk [RR]=1.02; 95% confidence interval [CI], 0.84-1.25) with regular aspirin use (325 mg every other day for 5 years or 100 mg every other day for 10 years). Six cohort studies that followed a total of 231,252 patients did report a decrease in CRC incidence over 4 to 10 years (RR=0.78; 95% CI, 0.63-0.97).¹

In a pooled analysis evaluating 2 primary prevention RCTs (the British Doctors Aspirin Trial and UK-TIA Aspirin Trial, total N=7588), aspirin was found to reduce the incidence of colorectal cancer (hazard ratio [HR]=0.74; 95% CI, 0.56-0.97; P=.02 overall; for aspirin given for 5 years or longer, HR=0.63; 95% CI, 0.47-0.85; P=.002). The effect was significant only at 10 to 14 years of follow-up (0 to 9 years: HR=0.92, 95% CI, 0.56-1.49, P=.73; 5 to 9 years: HR=1.08, 95% CI, 0.55-2.14, P=.83; 10 to 14 years: HR=0.51, 95% CI, 0.29-0.90, P=.02; 15 to 19 years: HR=0.70, 95% CI, 0.43-1.14, P=.15; ≥20 years: HR=0.90, 95% CI, 0.42-1.95, P=.79).²

Adverse effects, including stroke, are dose-dependent
The USPSTF review also summarized the harms associated with aspirin use. When aspirin was given for secondary prevention of stroke, the risk of hemorrhagic stroke was dose-dependent, varying from 0.3% to 1.1% (100 mg/d: 0.3%, 95% CI, 0.2%-0.4%; 100-325 mg/d: 0.3%, 95% CI, 0.2%-0.3%; 325 mg/d: 1.1%, 95% CI, 0.7%-1.5%).

Aspirin also was associated with an increased risk of gastrointestinal (GI) symptoms (odds ratio [OR]=1.7; 95% CI, 1.5-1.8), GI bleeding (RR=1.6-2.5), and hospitalization for GI bleeding (OR=1.9; 95% CI, 1.1-3.1). The risks of GI bleeding or perforation were dose-dependent.¹

Low-dose aspirin promotes secondary prevention of adenomas
In a Cochrane review evaluating the effects of aspirin on CRA, pooled data from 3 RCTs with a total of 1839 subjects (1322 with a history of CRA and 517 with a history of CRC) showed that aspirin in a daily dose of 81 mg is effective for secondary prevention of sporadic CRA over a 1- to 3-year follow-up period (RR=0.77; 95% CI, 0.61-0.96; number needed to treat=12.5). The outcome measured in these 3 trials was an intermediate clinical finding, CRA, and not the more relevant end point of CRC.³

Recommendations

The USPSTF recommends against routine use of aspirin and nonsteroidal anti-inflammatory drugs to prevent colorectal cancer in people at average risk (grade D recommendation: ineffective or harm outweighs benefits).⁴

The American Gastroenterological Association (AGA) doesn’t recommend aspirin for primary CRC prevention, but acknowledges a possible role in secondary prevention. Aspirin should be considered for patients with a personal history of CRC, advanced CRA, or a strong family history but no history of peptic ulcer disease or hemorrhagic stroke. The AGA notes that 1 in 100 people taking aspirin for 2 years will develop significant GI bleeding.⁵

Evidence summary

A systematic review conducted for the US Preventive Services Task Force (USPSTF) addressed the use of aspirin for primary prevention of colorectal carcinomas (CRC) and colorectal adenomas (CRA).

Pooled data from 2 randomized-controlled trials (RCTs) with a total of 61,947 patients showed no decrease in CRC incidence (relative risk [RR]=1.02; 95% confidence interval [CI], 0.84-1.25) with regular aspirin use (325 mg every other day for 5 years or 100 mg every other day for 10 years). Six cohort studies that followed a total of 231,252 patients did report a decrease in CRC incidence over 4 to 10 years (RR=0.78; 95% CI, 0.63-0.97).¹

In a pooled analysis evaluating 2 primary prevention RCTs (the British Doctors Aspirin Trial and UK-TIA Aspirin Trial, total N=7588), aspirin was found to reduce the incidence of colorectal cancer (hazard ratio [HR]=0.74; 95% CI, 0.56-0.97; P=.02 overall; for aspirin given for 5 years or longer, HR=0.63; 95% CI, 0.47-0.85; P=.002). The effect was significant only at 10 to 14 years of follow-up (0 to 9 years: HR=0.92, 95% CI, 0.56-1.49, P=.73; 5 to 9 years: HR=1.08, 95% CI, 0.55-2.14, P=.83; 10 to 14 years: HR=0.51, 95% CI, 0.29-0.90, P=.02; 15 to 19 years: HR=0.70, 95% CI, 0.43-1.14, P=.15; ≥20 years: HR=0.90, 95% CI, 0.42-1.95, P=.79).²

Adverse effects, including stroke, are dose-dependent
The USPSTF review also summarized the harms associated with aspirin use. When aspirin was given for secondary prevention of stroke, the risk of hemorrhagic stroke was dose-dependent, varying from 0.3% to 1.1% (100 mg/d: 0.3%, 95% CI, 0.2%-0.4%; 100-325 mg/d: 0.3%, 95% CI, 0.2%-0.3%; 325 mg/d: 1.1%, 95% CI, 0.7%-1.5%).

Aspirin also was associated with an increased risk of gastrointestinal (GI) symptoms (odds ratio [OR]=1.7; 95% CI, 1.5-1.8), GI bleeding (RR=1.6-2.5), and hospitalization for GI bleeding (OR=1.9; 95% CI, 1.1-3.1). The risks of GI bleeding or perforation were dose-dependent.¹

Low-dose aspirin promotes secondary prevention of adenomas
In a Cochrane review evaluating the effects of aspirin on CRA, pooled data from 3 RCTs with a total of 1839 subjects (1322 with a history of CRA and 517 with a history of CRC) showed that aspirin in a daily dose of 81 mg is effective for secondary prevention of sporadic CRA over a 1- to 3-year follow-up period (RR=0.77; 95% CI, 0.61-0.96; number needed to treat=12.5). The outcome measured in these 3 trials was an intermediate clinical finding, CRA, and not the more relevant end point of CRC.³

Recommendations

The USPSTF recommends against routine use of aspirin and nonsteroidal anti-inflammatory drugs to prevent colorectal cancer in people at average risk (grade D recommendation: ineffective or harm outweighs benefits).⁴

The American Gastroenterological Association (AGA) doesn’t recommend aspirin for primary CRC prevention, but acknowledges a possible role in secondary prevention. Aspirin should be considered for patients with a personal history of CRC, advanced CRA, or a strong family history but no history of peptic ulcer disease or hemorrhagic stroke. The AGA notes that 1 in 100 people taking aspirin for 2 years will develop significant GI bleeding.⁵

Evidence summary

A systematic review conducted for the US Preventive Services Task Force (USPSTF) addressed the use of aspirin for primary prevention of colorectal carcinomas (CRC) and colorectal adenomas (CRA).

Pooled data from 2 randomized-controlled trials (RCTs) with a total of 61,947 patients showed no decrease in CRC incidence (relative risk [RR]=1.02; 95% confidence interval [CI], 0.84-1.25) with regular aspirin use (325 mg every other day for 5 years or 100 mg every other day for 10 years). Six cohort studies that followed a total of 231,252 patients did report a decrease in CRC incidence over 4 to 10 years (RR=0.78; 95% CI, 0.63-0.97).¹

In a pooled analysis evaluating 2 primary prevention RCTs (the British Doctors Aspirin Trial and UK-TIA Aspirin Trial, total N=7588), aspirin was found to reduce the incidence of colorectal cancer (hazard ratio [HR]=0.74; 95% CI, 0.56-0.97; P=.02 overall; for aspirin given for 5 years or longer, HR=0.63; 95% CI, 0.47-0.85; P=.002). The effect was significant only at 10 to 14 years of follow-up (0 to 9 years: HR=0.92, 95% CI, 0.56-1.49, P=.73; 5 to 9 years: HR=1.08, 95% CI, 0.55-2.14, P=.83; 10 to 14 years: HR=0.51, 95% CI, 0.29-0.90, P=.02; 15 to 19 years: HR=0.70, 95% CI, 0.43-1.14, P=.15; ≥20 years: HR=0.90, 95% CI, 0.42-1.95, P=.79).²

Adverse effects, including stroke, are dose-dependent
The USPSTF review also summarized the harms associated with aspirin use. When aspirin was given for secondary prevention of stroke, the risk of hemorrhagic stroke was dose-dependent, varying from 0.3% to 1.1% (100 mg/d: 0.3%, 95% CI, 0.2%-0.4%; 100-325 mg/d: 0.3%, 95% CI, 0.2%-0.3%; 325 mg/d: 1.1%, 95% CI, 0.7%-1.5%).

Aspirin also was associated with an increased risk of gastrointestinal (GI) symptoms (odds ratio [OR]=1.7; 95% CI, 1.5-1.8), GI bleeding (RR=1.6-2.5), and hospitalization for GI bleeding (OR=1.9; 95% CI, 1.1-3.1). The risks of GI bleeding or perforation were dose-dependent.¹

Low-dose aspirin promotes secondary prevention of adenomas
In a Cochrane review evaluating the effects of aspirin on CRA, pooled data from 3 RCTs with a total of 1839 subjects (1322 with a history of CRA and 517 with a history of CRC) showed that aspirin in a daily dose of 81 mg is effective for secondary prevention of sporadic CRA over a 1- to 3-year follow-up period (RR=0.77; 95% CI, 0.61-0.96; number needed to treat=12.5). The outcome measured in these 3 trials was an intermediate clinical finding, CRA, and not the more relevant end point of CRC.³

Recommendations

The USPSTF recommends against routine use of aspirin and nonsteroidal anti-inflammatory drugs to prevent colorectal cancer in people at average risk (grade D recommendation: ineffective or harm outweighs benefits).⁴

The American Gastroenterological Association (AGA) doesn’t recommend aspirin for primary CRC prevention, but acknowledges a possible role in secondary prevention. Aspirin should be considered for patients with a personal history of CRC, advanced CRA, or a strong family history but no history of peptic ulcer disease or hemorrhagic stroke. The AGA notes that 1 in 100 people taking aspirin for 2 years will develop significant GI bleeding.⁵

Evidence summary

Otitis media is the most common outpatient diagnosis in children.³ Although these infections usually resolve without treatment, it is common practice in the United States to prescribe antibiotics.⁴

Antibiotic benefits are small in uncomplicated disease

A Cochrane review of 8 randomized controlled trials (RCTs)—6 double-blinded, 2287 children total—compared antibiotics with placebo for uncomplicated AOM in otherwise healthy children.⁵ The review showed that children treated with antibiotics were no less likely to have pain at 24 hours after starting therapy than untreated children. However, 7% fewer children who received antibiotics had pain at 2 to 7 days than unmedicated children (number needed to treat [NNT]=15; 95% confidence interval [CI], 11-24).

Children treated with antibiotics had no significant decrease in recurrence of AOM (mean 0.70 vs 0.63; 95% CI, -0.22 to 0.34; odds ratio [OR]=0.99) or hearing loss (reported by a combination of tympanometry and audiometry). They did have an increase in nausea, diarrhea, and rash, however. Only 1 case of mastoiditis was reported in the included studies.⁵

Of note, 2 of the studies showed a modest increase in failure rates of placebo treatment for children younger than 2 years and children with bilateral disease. Antibiotics may benefit these groups. Overall, both the potential benefits and harms of antibiotics for AOM are small.⁵

More on which children may benefit from antibiotics

A secondary analysis of cohorts from 6 RCTs (a total of 824 children untreated for AOM) identified age younger than 2 years (OR=2.07; 95% CI, 1.47-2.91; P<.0001) and bilateral disease (OR=1.70; 95% CI, 1.19-2.41; P=.003) as independent risk factors for pain and fever at 3 to 7 days of illness. However, the study did not address whether antibiotics would actually mitigate the risk factors.⁶

In another secondary analysis of cohorts from a single RCT (315 patients), children with high temperature or vomiting who were treated immediately with antibiotics were less likely to be in distress by day 3 of illness (32% immediate vs 53% delayed; P=.045; NNT=5) or have night disturbance (26% immediate vs 59% delayed; P=.002; NNT=3). The greatest benefit occurred among children younger than 2 years and children with bilateral infection (NNT=4). The outcomes were reported by parents, who were not blinded.⁷

One regimen is as effective as another

In meta-analyses of subsets of a systematic review that included 74 RCTs and 6 cohort studies, the 1- to 7-day clinical failure rate among children not given antibiotics for AOM was 19% (95% CI, 0.10-0.28). Patients treated with ampicillin or amoxicillin had a 2- to 7-dayclinical failure rate of 7% (NNT=8; 95% CI, 0.04-0.20). Pooled analyses did not show any difference in efficacy between comparisons of penicillin, ampicillin, amoxicillin (2 or 3 times daily; standard or high dose), amoxicillin-clavulanate, cefaclor, cefixime, ceftriaxone, azithromycin, and trimethoprim-sulfamethoxazole.

An 8% higher incidence of diarrhea was noted for cefixime compared with amoxicillin (number needed to harm=12; 95% CI, 0.04-0.13). Azithromycin had a 19% lower adverse event rate than amoxicillin-clavulanate (NNT=5; 95% CI, 0.09-0.29).⁸

Five days of treatment are as good as 10

A 2000 Cochrane review found that 5 days of antibiotic therapy was as effective as a 10-day course of treatment in otherwise healthy children with uncomplicated AOM. A slight increase in signs, symptoms, relapse, or reinfection among children receiving 5 days of antibiotics was noted at 8 to 18 days after treatment (OR=1.52; 95% CI, 1.17-1.98), but the finding was no longer statistically significant at 30 days (OR=1.22; 95% CI, 0.98-1.54).⁹

Recommendations

The American Academy of Pediatrics recommends amoxicillin 80 to 90 mg/kg per day in 2 divided doses for:

• all children younger than 6 months with AOM
• children 6 to 24 months old with a certain diagnosis of AOM (rapid onset, signs of middle-ear effusion, and signs and symptoms of middle-ear inflammation) or severe illness (moderate to severe otalgia or fever ≥102.2°F [39°C])
• children older than 24 months with severe illness.

All other children may be observed if the caregiver consents and is able to monitor the child and if systems are in place for follow-up communication, reevaluation, or access to medication.

Children with a non-type-I penicillin allergy can be given a second- or third-generation cephalosporin, such as cefdinir (14 mg/kg per day in 1 or 2 doses), cefpodoxime (10 mg/kg per day in 1 dose), or cefuroxime (30 mg/kg per day in 2 divided doses). If the child is at high risk of anaphylaxis, 2 acceptable options are azithromycin (10 mg/kg on day 1 followed by 5 mg/kg per day for 4 days as a single daily dose) or clarithromycin (15 mg/kg per day in 2 divided doses).

Amoxicillin should not be given to children at risk for highly amoxicillin-resistant organisms (eg, children who have had antibiotics in the previous 30 days, concomitant purulent conjunctivitis, chronic prophylactic amoxicillin). The recommended alternative is high-dose amoxicillin-clavulanate (90 mg/kg per day of amoxicillin and 6.4 mg/kg per day of clavulanate in 2 divided doses).⁴

Evidence summary

Otitis media is the most common outpatient diagnosis in children.³ Although these infections usually resolve without treatment, it is common practice in the United States to prescribe antibiotics.⁴

Antibiotic benefits are small in uncomplicated disease

A Cochrane review of 8 randomized controlled trials (RCTs)—6 double-blinded, 2287 children total—compared antibiotics with placebo for uncomplicated AOM in otherwise healthy children.⁵ The review showed that children treated with antibiotics were no less likely to have pain at 24 hours after starting therapy than untreated children. However, 7% fewer children who received antibiotics had pain at 2 to 7 days than unmedicated children (number needed to treat [NNT]=15; 95% confidence interval [CI], 11-24).

Children treated with antibiotics had no significant decrease in recurrence of AOM (mean 0.70 vs 0.63; 95% CI, -0.22 to 0.34; odds ratio [OR]=0.99) or hearing loss (reported by a combination of tympanometry and audiometry). They did have an increase in nausea, diarrhea, and rash, however. Only 1 case of mastoiditis was reported in the included studies.⁵

Of note, 2 of the studies showed a modest increase in failure rates of placebo treatment for children younger than 2 years and children with bilateral disease. Antibiotics may benefit these groups. Overall, both the potential benefits and harms of antibiotics for AOM are small.⁵

More on which children may benefit from antibiotics

A secondary analysis of cohorts from 6 RCTs (a total of 824 children untreated for AOM) identified age younger than 2 years (OR=2.07; 95% CI, 1.47-2.91; P<.0001) and bilateral disease (OR=1.70; 95% CI, 1.19-2.41; P=.003) as independent risk factors for pain and fever at 3 to 7 days of illness. However, the study did not address whether antibiotics would actually mitigate the risk factors.⁶

In another secondary analysis of cohorts from a single RCT (315 patients), children with high temperature or vomiting who were treated immediately with antibiotics were less likely to be in distress by day 3 of illness (32% immediate vs 53% delayed; P=.045; NNT=5) or have night disturbance (26% immediate vs 59% delayed; P=.002; NNT=3). The greatest benefit occurred among children younger than 2 years and children with bilateral infection (NNT=4). The outcomes were reported by parents, who were not blinded.⁷

One regimen is as effective as another

In meta-analyses of subsets of a systematic review that included 74 RCTs and 6 cohort studies, the 1- to 7-day clinical failure rate among children not given antibiotics for AOM was 19% (95% CI, 0.10-0.28). Patients treated with ampicillin or amoxicillin had a 2- to 7-dayclinical failure rate of 7% (NNT=8; 95% CI, 0.04-0.20). Pooled analyses did not show any difference in efficacy between comparisons of penicillin, ampicillin, amoxicillin (2 or 3 times daily; standard or high dose), amoxicillin-clavulanate, cefaclor, cefixime, ceftriaxone, azithromycin, and trimethoprim-sulfamethoxazole.

An 8% higher incidence of diarrhea was noted for cefixime compared with amoxicillin (number needed to harm=12; 95% CI, 0.04-0.13). Azithromycin had a 19% lower adverse event rate than amoxicillin-clavulanate (NNT=5; 95% CI, 0.09-0.29).⁸

Five days of treatment are as good as 10

A 2000 Cochrane review found that 5 days of antibiotic therapy was as effective as a 10-day course of treatment in otherwise healthy children with uncomplicated AOM. A slight increase in signs, symptoms, relapse, or reinfection among children receiving 5 days of antibiotics was noted at 8 to 18 days after treatment (OR=1.52; 95% CI, 1.17-1.98), but the finding was no longer statistically significant at 30 days (OR=1.22; 95% CI, 0.98-1.54).⁹

Recommendations

The American Academy of Pediatrics recommends amoxicillin 80 to 90 mg/kg per day in 2 divided doses for:

• all children younger than 6 months with AOM
• children 6 to 24 months old with a certain diagnosis of AOM (rapid onset, signs of middle-ear effusion, and signs and symptoms of middle-ear inflammation) or severe illness (moderate to severe otalgia or fever ≥102.2°F [39°C])
• children older than 24 months with severe illness.

All other children may be observed if the caregiver consents and is able to monitor the child and if systems are in place for follow-up communication, reevaluation, or access to medication.

Children with a non-type-I penicillin allergy can be given a second- or third-generation cephalosporin, such as cefdinir (14 mg/kg per day in 1 or 2 doses), cefpodoxime (10 mg/kg per day in 1 dose), or cefuroxime (30 mg/kg per day in 2 divided doses). If the child is at high risk of anaphylaxis, 2 acceptable options are azithromycin (10 mg/kg on day 1 followed by 5 mg/kg per day for 4 days as a single daily dose) or clarithromycin (15 mg/kg per day in 2 divided doses).

Amoxicillin should not be given to children at risk for highly amoxicillin-resistant organisms (eg, children who have had antibiotics in the previous 30 days, concomitant purulent conjunctivitis, chronic prophylactic amoxicillin). The recommended alternative is high-dose amoxicillin-clavulanate (90 mg/kg per day of amoxicillin and 6.4 mg/kg per day of clavulanate in 2 divided doses).⁴

Evidence summary

Otitis media is the most common outpatient diagnosis in children.³ Although these infections usually resolve without treatment, it is common practice in the United States to prescribe antibiotics.⁴

Antibiotic benefits are small in uncomplicated disease

A Cochrane review of 8 randomized controlled trials (RCTs)—6 double-blinded, 2287 children total—compared antibiotics with placebo for uncomplicated AOM in otherwise healthy children.⁵ The review showed that children treated with antibiotics were no less likely to have pain at 24 hours after starting therapy than untreated children. However, 7% fewer children who received antibiotics had pain at 2 to 7 days than unmedicated children (number needed to treat [NNT]=15; 95% confidence interval [CI], 11-24).

Children treated with antibiotics had no significant decrease in recurrence of AOM (mean 0.70 vs 0.63; 95% CI, -0.22 to 0.34; odds ratio [OR]=0.99) or hearing loss (reported by a combination of tympanometry and audiometry). They did have an increase in nausea, diarrhea, and rash, however. Only 1 case of mastoiditis was reported in the included studies.⁵

Of note, 2 of the studies showed a modest increase in failure rates of placebo treatment for children younger than 2 years and children with bilateral disease. Antibiotics may benefit these groups. Overall, both the potential benefits and harms of antibiotics for AOM are small.⁵

More on which children may benefit from antibiotics

A secondary analysis of cohorts from 6 RCTs (a total of 824 children untreated for AOM) identified age younger than 2 years (OR=2.07; 95% CI, 1.47-2.91; P<.0001) and bilateral disease (OR=1.70; 95% CI, 1.19-2.41; P=.003) as independent risk factors for pain and fever at 3 to 7 days of illness. However, the study did not address whether antibiotics would actually mitigate the risk factors.⁶

In another secondary analysis of cohorts from a single RCT (315 patients), children with high temperature or vomiting who were treated immediately with antibiotics were less likely to be in distress by day 3 of illness (32% immediate vs 53% delayed; P=.045; NNT=5) or have night disturbance (26% immediate vs 59% delayed; P=.002; NNT=3). The greatest benefit occurred among children younger than 2 years and children with bilateral infection (NNT=4). The outcomes were reported by parents, who were not blinded.⁷

One regimen is as effective as another

In meta-analyses of subsets of a systematic review that included 74 RCTs and 6 cohort studies, the 1- to 7-day clinical failure rate among children not given antibiotics for AOM was 19% (95% CI, 0.10-0.28). Patients treated with ampicillin or amoxicillin had a 2- to 7-dayclinical failure rate of 7% (NNT=8; 95% CI, 0.04-0.20). Pooled analyses did not show any difference in efficacy between comparisons of penicillin, ampicillin, amoxicillin (2 or 3 times daily; standard or high dose), amoxicillin-clavulanate, cefaclor, cefixime, ceftriaxone, azithromycin, and trimethoprim-sulfamethoxazole.

An 8% higher incidence of diarrhea was noted for cefixime compared with amoxicillin (number needed to harm=12; 95% CI, 0.04-0.13). Azithromycin had a 19% lower adverse event rate than amoxicillin-clavulanate (NNT=5; 95% CI, 0.09-0.29).⁸

Five days of treatment are as good as 10

A 2000 Cochrane review found that 5 days of antibiotic therapy was as effective as a 10-day course of treatment in otherwise healthy children with uncomplicated AOM. A slight increase in signs, symptoms, relapse, or reinfection among children receiving 5 days of antibiotics was noted at 8 to 18 days after treatment (OR=1.52; 95% CI, 1.17-1.98), but the finding was no longer statistically significant at 30 days (OR=1.22; 95% CI, 0.98-1.54).⁹

Recommendations

The American Academy of Pediatrics recommends amoxicillin 80 to 90 mg/kg per day in 2 divided doses for:

• all children younger than 6 months with AOM
• children 6 to 24 months old with a certain diagnosis of AOM (rapid onset, signs of middle-ear effusion, and signs and symptoms of middle-ear inflammation) or severe illness (moderate to severe otalgia or fever ≥102.2°F [39°C])
• children older than 24 months with severe illness.

All other children may be observed if the caregiver consents and is able to monitor the child and if systems are in place for follow-up communication, reevaluation, or access to medication.

Children with a non-type-I penicillin allergy can be given a second- or third-generation cephalosporin, such as cefdinir (14 mg/kg per day in 1 or 2 doses), cefpodoxime (10 mg/kg per day in 1 dose), or cefuroxime (30 mg/kg per day in 2 divided doses). If the child is at high risk of anaphylaxis, 2 acceptable options are azithromycin (10 mg/kg on day 1 followed by 5 mg/kg per day for 4 days as a single daily dose) or clarithromycin (15 mg/kg per day in 2 divided doses).

Amoxicillin should not be given to children at risk for highly amoxicillin-resistant organisms (eg, children who have had antibiotics in the previous 30 days, concomitant purulent conjunctivitis, chronic prophylactic amoxicillin). The recommended alternative is high-dose amoxicillin-clavulanate (90 mg/kg per day of amoxicillin and 6.4 mg/kg per day of clavulanate in 2 divided doses).⁴

Recommendations from others

Three primary care textbooks recommend providing patients with simple written instructions to enhance compliance with medications.^1-3 Another text⁴ also advocates simple written instructions and recommends follow-up reminders, addressing compliance issues at patient visits, and maintaining compliance interventions throughout the duration of treatment.

Evidence summary

Medication adherence is defined as the extent to which patients follow instructions for prescribed treatments.⁵ Estimates of medication adherence range from 0% to 100% and are typically approximately 50%.⁵ The Cochrane Database of Systematic Reviews evaluated the published, unbiased, randomized controlled trials of interventions to improve medication adherence that measured both adherence and treatment outcome. Eighty percent follow-up was required, as well as a 6-month follow-up period for long-term treatments with initial positive outcomes. A total of 4762 citations were evaluated; only 19 met their criteria for full consideration. We did not find any more recent or additional trials that met the Cochrane criteria for consideration.

The interventions studied included combinations of patient and family counseling (verbal, written, telephone, and automated); patient involvement in self-monitoring of outcomes (eg, blood pressure); simplified medication regimens; reminders of appointments, refills, and pill schedules; and care provided at the workplace. All interventions were for single clinical conditions including hypertension (7 studies), schizophrenia and psychosis (6), asthma (2), acute infection (2), hyperlipidemia (1), and epilepsy (1). Both of the studies concerning simplifying medication regimens showed a positive effect on medication adherence. Nine of 16 studies on varying strategies of interventions reported statistically significant increases in medication adherence. Most of the studies that used multidimensional interventions did not assess the individual components of the interventions; therefore, it is impossible to determine if one part of a complex intervention was more important than another. Also, only small improvements in adherence or treatment outcomes were observed in these studies.

This review was designed to specifically address interventions to improve patient compliance with prescribed medication. Other areas of potential significance to patient compliance are health care access, continuity of care, and medication side effects. In addition to simplifying dosing regimens, it would stand to reason that maximizing the provision of care and minimizing adverse medication side effects would have a positive effect on patient compliance.

Recommendations from others

Three primary care textbooks recommend providing patients with simple written instructions to enhance compliance with medications.^1-3 Another text⁴ also advocates simple written instructions and recommends follow-up reminders, addressing compliance issues at patient visits, and maintaining compliance interventions throughout the duration of treatment.

Evidence summary

Medication adherence is defined as the extent to which patients follow instructions for prescribed treatments.⁵ Estimates of medication adherence range from 0% to 100% and are typically approximately 50%.⁵ The Cochrane Database of Systematic Reviews evaluated the published, unbiased, randomized controlled trials of interventions to improve medication adherence that measured both adherence and treatment outcome. Eighty percent follow-up was required, as well as a 6-month follow-up period for long-term treatments with initial positive outcomes. A total of 4762 citations were evaluated; only 19 met their criteria for full consideration. We did not find any more recent or additional trials that met the Cochrane criteria for consideration.

The interventions studied included combinations of patient and family counseling (verbal, written, telephone, and automated); patient involvement in self-monitoring of outcomes (eg, blood pressure); simplified medication regimens; reminders of appointments, refills, and pill schedules; and care provided at the workplace. All interventions were for single clinical conditions including hypertension (7 studies), schizophrenia and psychosis (6), asthma (2), acute infection (2), hyperlipidemia (1), and epilepsy (1). Both of the studies concerning simplifying medication regimens showed a positive effect on medication adherence. Nine of 16 studies on varying strategies of interventions reported statistically significant increases in medication adherence. Most of the studies that used multidimensional interventions did not assess the individual components of the interventions; therefore, it is impossible to determine if one part of a complex intervention was more important than another. Also, only small improvements in adherence or treatment outcomes were observed in these studies.

This review was designed to specifically address interventions to improve patient compliance with prescribed medication. Other areas of potential significance to patient compliance are health care access, continuity of care, and medication side effects. In addition to simplifying dosing regimens, it would stand to reason that maximizing the provision of care and minimizing adverse medication side effects would have a positive effect on patient compliance.

Recommendations from others

Three primary care textbooks recommend providing patients with simple written instructions to enhance compliance with medications.^1-3 Another text⁴ also advocates simple written instructions and recommends follow-up reminders, addressing compliance issues at patient visits, and maintaining compliance interventions throughout the duration of treatment.

Evidence summary

Medication adherence is defined as the extent to which patients follow instructions for prescribed treatments.⁵ Estimates of medication adherence range from 0% to 100% and are typically approximately 50%.⁵ The Cochrane Database of Systematic Reviews evaluated the published, unbiased, randomized controlled trials of interventions to improve medication adherence that measured both adherence and treatment outcome. Eighty percent follow-up was required, as well as a 6-month follow-up period for long-term treatments with initial positive outcomes. A total of 4762 citations were evaluated; only 19 met their criteria for full consideration. We did not find any more recent or additional trials that met the Cochrane criteria for consideration.

The interventions studied included combinations of patient and family counseling (verbal, written, telephone, and automated); patient involvement in self-monitoring of outcomes (eg, blood pressure); simplified medication regimens; reminders of appointments, refills, and pill schedules; and care provided at the workplace. All interventions were for single clinical conditions including hypertension (7 studies), schizophrenia and psychosis (6), asthma (2), acute infection (2), hyperlipidemia (1), and epilepsy (1). Both of the studies concerning simplifying medication regimens showed a positive effect on medication adherence. Nine of 16 studies on varying strategies of interventions reported statistically significant increases in medication adherence. Most of the studies that used multidimensional interventions did not assess the individual components of the interventions; therefore, it is impossible to determine if one part of a complex intervention was more important than another. Also, only small improvements in adherence or treatment outcomes were observed in these studies.

This review was designed to specifically address interventions to improve patient compliance with prescribed medication. Other areas of potential significance to patient compliance are health care access, continuity of care, and medication side effects. In addition to simplifying dosing regimens, it would stand to reason that maximizing the provision of care and minimizing adverse medication side effects would have a positive effect on patient compliance.