Abstract
Background
Studies among adults have found that physicians who are high antibiotic prescribers are also high prescribers of other medications that are frequently overprescribed or prescribed inappropriately. The extent to which these findings generalise to prescribing for paediatric populations is unknown.
Methods
Population-based cohort study, Ontario, Canada, involving child-physician pairs with ≥1 primary care visit, 2018–2019. We used multivariable, hierarchical logistic regression to determine the association between high-intensity antibiotic prescribing and prescribing of stimulants, antidepressants, antipsychotics or all three drug classes, adjusting for child characteristics and clustering of child-physician pairs within physicians.
Results
2 531 392 visits were included (834 565 children, 10 286 physicians). Psychotropic drug, stimulant and antidepressant prescribing was lower among high antibiotic prescribers versus average/low-intensity prescribers (adjusted OR (aOR) 0.66 (95% CI 0.62 to 0.71), 0.67 (95% CI 0.62 to 0.72) and 0.79 (95% CI 0.70 to 0.89), respectively). There was no association between the odds of a physician prescribing an antipsychotic and high-intensity antibiotic prescribing (aOR 0.95 (95% CI 0.84 to 1.08)). The inverse association between prescribing of psychotropic drugs and high-intensity antibiotic prescribing was stronger for visits with children with high continuity of care versus average/low continuity of care (aOR 0.42 (95% CI 0.34 to 0.52) vs aOR 0.72 (95% CI 0.66 to 78)), visits with children’s rostered versus non-rostered primary care physicians (aOR 0.54 (95% CI 0.48 to 0.59) vs aOR 0.74 (95% CI 0.68 to 0.81)) and visits with physicians remunerated under a fee-for-service versus salaried/capitation-based models (aOR 0.63 (95% CI 0.56 to 0.70) vs aOR 0.86 (95% CI 0.77 to 0.97)).
Conclusions
High antibiotic prescribing was generally associated with a lower propensity to prescribe psychotropic drugs among children. Our findings suggest the need for tailored approaches to paediatric drug stewardship across different classes of drugs.
Keywords: Child Psychiatry, Health services research, Child Health
WHAT IS ALREADY KNOWN ON THIS TOPIC
Studies examining physician prescribing practices for adults have found that physicians who are high-volume antibiotic prescribers are also high prescribers of other medications that are frequently overprescribed or prescribed inappropriately (eg, opioids, benzodiazepines, antipsychotics), suggesting that broad stewardship interventions across drug classes might be an efficient way to reduce overprescribing. The extent to which these findings apply to prescribing for children is unknown.
WHAT THIS STUDY ADDS
Contrary to our expectations, we found that physicians who prescribed high volumes of antibiotics for children were less likely to prescribe psychotropic drugs. We also found that factors such as care continuity and physician reimbursement model influenced prescribing behaviour in both directions.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Our results do not support a ‘one-size-fits-all’ approach to paediatric drug stewardship but instead suggest the need for tailored approaches across different classes of drugs.
Introduction
Prescriptions for stimulants, antidepressants and antipsychotics among children have steadily increased in Canada1,3 and multiple countries worldwide,3,5 despite ongoing controversy about safety and effectiveness6,9 and concerns about overdiagnosis and overprescribing.10 Likewise, children are often prescribed antibiotics inappropriately,11 12 which contributes to rising antibiotic resistance13 and can lead to adverse drug effects, including allergic reactions,14 organ toxicities15 and Clostridioides difficile-associated diarrhoea.14 16
Prescribing rates for psychotropic drugs17 and antibiotics18 are highly variable. This variability is not fully explained by practice setting or patient-specific clinical or sociodemographic factors, suggesting that physician prescribing habits and preferences might be important drivers of variability. Studies examining physician prescribing behaviour for adults and nursing home residents have found that those who are high-antibiotic prescribers are also high prescribers of other medications that are frequently overprescribed or prescribed inappropriately (eg, opioids, benzodiazepines, antipsychotics).19,22 The extent to which these findings generalise to prescribing for paediatric populations is unknown.
In Canada, children can receive psychotropic drug prescriptions from primary care physicians, paediatricians or psychiatrists, with studies indicating that primary care physicians and paediatricians write the majority of these prescriptions.23 24 Psychiatrists are often involved when cases are more complex or when a child has co-occurring psychiatric conditions. Understanding the association, if any, between high antibiotic prescribing and prescribing of psychotropic drugs among children could inform the development of interventions to improve prescribing of multiple drug classes by primary care physicians and paediatricians in this population. As such, our aim was to evaluate the association between physicians’ antibiotic prescribing intensity and their propensity to prescribe stimulants, antidepressants and antipsychotics for children. We also examined the characteristics of physicians and the children they cared for by physician antibiotic and psychotropic drug prescribing intensity.
Methods
Setting
Ontario is Canada’s most populous province with approximately 14.6 million residents (2019), including 1.1 million residents aged 6–12 years. The provincial universal healthcare system provides physician and hospital services at no direct cost to all residents. In January 2018, the Ontario government implemented a publicly funded Pharmacare programme, which provided prescription drugs covered under the Ontario Drug Benefit (ODB) programme at no cost to residents aged <25 years. In April 2019, the programme was changed to only cover drugs for children and young people without private drug insurance. The period of universal Pharmacare (ie, from January 2018 to March 2019) was a unique opportunity to examine antibiotic and psychotropic drug prescribing among children without potential selection bias due to variability in drug insurance.
Population
The source population included all eligible child-physician pairs with ≥1 primary care visit from 7 January 2018 to 3 March 2019. Eligible children were aged 6–12 years and enrolled in the Ontario Health Insurance Plan (OHIP). Eligible physicians were certified in family medicine or paediatrics and had billed ≥20 primary care visits for children aged 6–12 years during the study.
Data sources
We extracted prescription data from the ODB database, which contains comprehensive records of all publicly funded drugs dispensed to Ontario residents at outpatient pharmacies. Concordance between the ODB database and pharmacy records is >99%.25 We extracted physician demographics, medical training history and practice characteristics from the Ontario Corporate Provider Database and the Institute for Clinical Evaluative Sciences (ICES) Physician Database. We used the Registered Persons Database, Ontario’s universal health insurance registry, to ascertain the age, sex and postal code of children on the date of their visit. We linked each child’s postal code to the Canadian Census to determine rurality and to ascertain area-level income and marginalisation across four dimensions: (1) material resources; (2) households and dwellings; (3) age and labour force and (4) racial and newcomer populations.26 We identified the presence of paediatric complex chronic conditions using a validated algorithm based on the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Canada, and Canadian Classification of Interventions codes in the Canadian Institute for Health Information’s Discharge Abstract Database and Same Day Surgery Database.27 We ascertained mental health-related outpatient service use ≤2 years before the visit date by applying a validated algorithm to OHIP physician billing data.28 We identified each child’s rostered primary care physician using information in the Client Agency Programme Enrolment Database and primary care fee codes in OHIP. For children who had >2 primary care visits during the study, we calculated their continuity of care index, which reflects the distribution of their primary care visits between different physicians.29 Values range from 0, which indicates a different physician was seen at each visit (minimum continuity), to 1, when the same physician was seen for all visits (maximum continuity).29 We determined children’s immigration status using the Ontario portion of the Immigration, Refugee and Citizenship Canada Permanent Resident Database.
Databases were linked with unique coded identifiers and analysed at ICES, an independent, non-profit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyse healthcare and demographic data, without consent, for health system evaluation and improvement.
Antibiotic and psychotropic drug prescribing
We included all systemic antibiotic prescriptions dispensed at outpatient pharmacies ≤7 days after an eligible primary care visit. We used each physician’s antibiotic initiation rate, calculated by summing their antibiotic prescriptions for children aged 6–12 years during the study period and dividing by their total number of visits for children aged 6–12 years during the study period, to quantify their antibiotic prescribing intensity. This metric is strongly correlated with unnecessary antibiotic prescribing among Ontario primary care physicians (Pearson’s r=0.93).30 We categorised physicians as ‘low’ (quartile one), ‘average’ (quartile two and three) or ‘high’ (quartile four) intensity antibiotic prescribers.31
We extracted prescription data for stimulants, antidepressants and antipsychotics (hereafter collectively referred to as ‘study psychotropic drugs’) dispensed ≤28 days after an eligible visit at an outpatient pharmacy, as well as refill prescriptions (ie, dispensed >28 days after an eligible visit and within the study period). We calculated each physician’s psychotropic drug prescribing rate as the total days of therapy (including refills) of the psychotropic drug class of interest prescribed for children aged 6–12 during the study, divided by the physician’s total number of visits for children aged 6–12 during the study. We used days of therapy as the numerator instead of the number of prescriptions because psychotropic drug prescription days’ supply was highly variable, with some physicians prescribing shorter durations more frequently and others prescribing longer durations less frequently. We categorised physicians as low, average or high-intensity psychotropic drug prescribers based on prescribing rate quartiles as described for antibiotics.
Statistical analyses
We described physicians and the children they cared for using counts (percentages), mean±SD or median (IQR). We assessed differences between physician antibiotic and psychotropic drug prescribing intensity groups using standardised mean differences, with average-intensity prescribers as the comparator group for low-intensity and high-intensity prescribing groups. We considered an absolute standardised mean difference <0.1 as indicative of balance between groups.32 We applied the Spearman correlation test to examine correlations between physicians’ antibiotic prescribing rate and their stimulant, antidepressant and antipsychotic prescribing rates.
We used multivariable, hierarchical logistic regression with a binomial distribution to determine the association between high-intensity antibiotic prescribing and the likelihood of prescribing a stimulant, antidepressant, antipsychotic or all three drug classes. The unit of analysis was the primary care visit. Psychotropic prescription refills were considered part of the index prescription for the purposes of this analysis. We adjusted for child characteristics (age, sex, paediatric chronic complex conditions, immigration status, visit with rostered primary care physician, rural residence and area-level income and marginalisation on the racial and newcomer populations dimension) and accounted for clustering of child-physician pairs within physicians. We excluded visits with missing covariates (0.5% of visits).
To reduce the computational burden of using the entire population, we used a case-cohort design for regression analyses.33 With this approach, a 10% random sample of eligible visits was drawn from the source population along with all cases within the source population. We defined cases as visits linked with a prescription for a psychotropic drug of interest ≤28 days after the visit date. We weighted each observation by its probability of being included as a case or selected in the random sample. We performed analyses using SAS, V.9.4, and the R package lme4, V.1.1-7.
Exploratory analyses
We performed several post hoc exploratory analyses. First, we examined whether results differed for visits with children who had a high continuity versus a low continuity of care index, which we defined as an index above or below the 75th percentile, respectively. Second, we repeated the analyses separately for visits with a child’s rostered primary care physician versus visits that were not. Third, we performed subgroup analyses by physician remuneration model (predominantly fee-for-service vs salaried or capitation-based models). Last, we examined whether results were altered after excluding outlier physicians whose antibiotic initiations were below the fifth percentile.
Patient and public involvement
We did not involve patients or the public in the design, conduct, reporting or dissemination plans of our study due to time limitations. It was not possible to inform study participants of the results because we used de-identified data. Results will be disseminated through manuscript publication and legacy and social media.
Results
Over the study period, 2 531 392 visits met the inclusion criteria (figure 1). Visits included 834 565 children cared for by 10 286 physicians (online supplemental figure S1). Nearly one in six visits (408 179; 16%) was associated with ≥1 antibiotic prescription; 136 905 (5.4%), 19 542 (0.8%) and 23 053 (0.9%) visits were linked with ≥1 stimulant, antidepressant or antipsychotic prescription, respectively.
Figure 1. Summary of children’s primary care visits and antibiotic and psychotropic drug prescriptions, January 2018 to March 2019, Ontario, Canada.
Characteristics of physicians and children by physician antibiotic prescribing intensity
The characteristics of physicians by antibiotic prescribing intensity group are shown in table 1. Compared with average-intensity antibiotic prescribers, high-intensity antibiotic prescribers had a higher proportion of physicians who were male, older, late-career (ie, practicing for >25 years), international medical school graduates, certified in family medicine (vs paediatrics) and predominantly remunerated under a fee-for-service model. They also had higher visit volumes for 6–12-year-olds. The characteristics of children cared for by low-intensity, average-intensity or high-intensity antibiotic prescribers were generally similar, except high-intensity antibiotic prescribers cared for a higher proportion of children who were not their rostered patients and had a lower continuity of care index (table 2).
Table 1. Characteristics of physicians by their antibiotic prescribing intensity, 2018–2019, Ontario.
| Physician characteristic | Physician antibiotic prescribing intensity* | Mean standardised difference† | |||
|---|---|---|---|---|---|
| Low N=2571 |
Average N=5143 |
High N=2572 |
Low versus average | High versus average | |
| Sex, n (%) | |||||
| Female | 1433 (55.7) | 2663 (51.8) | 998 (38.8) | 0.08 | 0.26 |
| Male | 1138 (44.3) | 2480 (48.2) | 1574 (61.2) | 0.08 | 0.26 |
| Age group, n (%) | |||||
| <45 years | 622 (24.2) | 1159 (22.5) | 440 (17.1) | 0.04 | 0.14 |
| 45–64 years | 1120 (43.6) | 2381 (46.3) | 1213 (47.2) | 0.05 | 0.02 |
| ≥65 years | 332 (12.9) | 564 (11.0) | 403 (15.7) | 0.06 | 0.14 |
| Missing | 497 (19.3) | 1039 (20.2) | 516 (20.1) | 0.02 | 0 |
| Years practising, n (%) | |||||
| ≤5 years | 279 (10.9) | 580 (11.3) | 206 (8.0) | 0.01 | 0.11 |
| 6–10 years | 400 (15.6) | 764 (14.9) | 326 (12.7) | 0.02 | 0.06 |
| 11–25 years | 819 (31.9) | 1559 (30.3) | 745 (29.0) | 0.03 | 0.03 |
| >25 years | 1061 (41.3) | 2212 (43.0) | 1282 (49.8) | 0.04 | 0.14 |
| Missing | 12 (0.5) | 28 (0.5) | 13 (0.5) | 0.01 | 0.01 |
| Main speciality, n (%) | |||||
| Family medicine | 2058 (80.0) | 4860 (94.5) | 2501 (97.2) | 0.44 | 0.14 |
| Paediatrics | 513 (20.0) | 283 (5.5) | 71 (2.8) | 0.44 | 0.14 |
| Country of medical training, n (%) | |||||
| Canada | 1518 (59.0) | 2783 (54.1) | 1205 (46.9) | 0.10 | 0.15 |
| USA | 23 (0.9) | 26 (0.5) | 17 (0.7) | 0.05 | 0.02 |
| International | 533 (20.7) | 1295 (25.2) | 834 (32.4) | 0.11 | 0.16 |
| Missing | 497 (19.3) | 1039 (20.2) | 516 (20.1) | 0.02 | 0 |
| Predominantly fee-for-service remuneration model, n (%) | |||||
| Yes | 1126 (43.8) | 2026 (39.4) | 1511 (58.7) | 0.09 | 0.39 |
| No | 1445 (56.2) | 3117 (60.6) | 1061 (41.3) | 0.09 | 0.39 |
| Rural practice setting, n (%) | 225 (8.8) | 402 (7.8) | 155 (6.0) | 0.03 | 0.07 |
| Median (IQR) number of visits among 6–12-year-olds during the study | 92 (51–186) | 120 (69–216) | 175 (86–431) | 0.25 | 0.36 |
| Mean±SD per cent of visits for 6–12-year-olds during the study | 9.0±13.4 | 4.7±5.6 | 5.1±4.2 | 0.41 | 0.06 |
Antibiotic initiation rate was defined as the sum of each physician’s antibiotic prescriptions for children aged 6–12 years during the study period divided by the total number of visits for children aged 6–12 years during the study period.
Mean standardised differences ≥0.10 indicate potentially clinically important differences between groups and are shown in bold.
n, number.
Table 2. Characteristics of child visits by physician antibiotic prescribing intensity, 2018–2019, Ontario.
| Child characteristics | Physician antibiotic prescribing intensity* | Mean standardised difference† | |||
|---|---|---|---|---|---|
| Low N=541 489 |
Average N=1 151 874 |
High N=838 029 |
Low versus average | High versus average | |
| Sex, n (%) | |||||
| Female | 240 922 (44.5) | 559 509 (48.6) | 414 404 (49.4) | 0.08 | 0.02 |
| Male | 300 567 (55.5) | 592 365 (51.4) | 423 625 (50.6) | 0.08 | 0.02 |
| Mean±SD age, years | 8.9±2.0 | 8.9±2.0 | 8.8±2.0 | 0.03 | 0.01 |
| Rural residence, n (%) | 40 531 (7.5) | 50 548 (4.4) | 29 341 (3.5) | 0.13 | 0.05 |
| Immigration status, n (%) | |||||
| Born in Canada | 513 878 (94.9) | 1 055 372 (91.6) | 761 251 (90.8) | 0.13 | 0.03 |
| Immigrant | 21 528 (4.0) | 72 159 (6.3) | 54 723 (6.5) | 0.10 | 0.01 |
| Refugee | 6083 (1.1) | 24 343 (2.1) | 22 055 (2.6) | 0.08 | 0.03 |
| Income quintile, n (%) | |||||
| 1 (lowest) | 95 091 (17.6) | 207 361 (18.0) | 156 851 (18.7) | 0.01 | 0.02 |
| 2 | 96 403 (17.8) | 200 949 (17.4) | 148 682 (17.7) | 0.01 | 0.01 |
| 3 | 106 502 (19.7) | 234 952 (20.4) | 179 735 (21.4) | 0.02 | 0.03 |
| 4 | 119 379 (22.0) | 256 295 (22.3) | 187 866 (22.4) | 0 | 0 |
| 5 (highest) | 122 703 (22.7) | 249 604 (21.7) | 162 612 (19.4) | 0.02 | 0.06 |
| Missing | 1411 (0.3) | 2713 (0.2) | 2283 (0.3) | 0.01 | 0.01 |
| Households and dwellings quintile, n (%) | |||||
| 1 (least marginalised) | 154 809 (28.6) | 371 964 (32.3) | 276 941 (33.0) | 0.08 | 0.02 |
| 2 | 106 839 (19.7) | 220 060 (19.1) | 156 845 (18.7) | 0.02 | 0.01 |
| 3 | 94 248 (17.4) | 185 460 (16.1) | 135 704 (16.2) | 0.03 | 0 |
| 4 | 92 105 (17.0) | 175 282 (15.2) | 125 329 (15.0) | 0.05 | 0.01 |
| 5 (most marginalised) | 89 126 (16.5) | 193 988 (16.8) | 139 193 (16.6) | 0.01 | 0.01 |
| Missing | 4362 (0.8) | 5120 (0.4) | 4017 (0.5) | 0.05 | 0.01 |
| Material resources quintile, n (%) | |||||
| 1 (least marginalised) | 130 919 (24.2) | 284 334 (24.7) | 179 247 (21.4) | 0.01 | 0.08 |
| 2 | 119 415 (22.1) | 254 419 (22.1) | 187 591 (22.4) | 0 | 0.01 |
| 3 | 99 478 (18.4) | 210 922 (18.3) | 163 718 (19.5) | 0 | 0.03 |
| 4 | 89 522 (16.5) | 186 656 (16.2) | 141 417 (16.9) | 0.01 | 0.02 |
| 5 (most marginalised) | 97 793 (18.1) | 210 423 (18.3) | 162 039 (19.3) | 0.01 | 0.03 |
| Missing | 4362 (0.8) | 5120 (0.4) | 4017 (0.5) | 0.05 | 0.01 |
| Racial and newcomer populations quintile, n (%) | |||||
| 1 (least marginalised) | 58 846 (10.9) | 82 231 (7.1) | 56 398 (6.7) | 0.13 | 0.02 |
| 2 | 73 112 (13.5) | 121 420 (10.5) | 89 420 (10.7) | 0.09 | 0 |
| 3 | 91 005 (16.8) | 173 059 (15.0) | 126 366 (15.1) | 0.05 | 0 |
| 4 | 120 982 (22.3) | 260 138 (22.6) | 190 684 (22.8) | 0.01 | 0 |
| 5 (most marginalised) | 193 182 (35.7) | 509 906 (44.3) | 371 144 (44.3) | 0.18 | 0 |
| Missing | 4362 (0.8) | 5120 (0.4) | 4017 (0.5) | 0.05 | 0.01 |
| Age and labour force quintile, n (%) | |||||
| 1 (least marginalised) | 187 792 (34.7) | 454 815 (39.5) | 345 524 (41.2) | 0.10 | 0.04 |
| 2 | 114 570 (21.2) | 253 608 (22.0) | 176 308 (21.0) | 0.02 | 0.02 |
| 3 | 88 380 (16.3) | 176 996 (15.4) | 126 319 (15.1) | 0.03 | 0.01 |
| 4 | 80 072 (14.8) | 150 388 (13.1) | 106 534 (12.7) | 0.05 | 0.01 |
| 5 (most marginalised) | 66 313 (12.2) | 110 947 (9.6) | 79 327 (9.5) | 0.08 | 0.01 |
| Missing | 4362 (0.8) | 5120 (0.4) | 4017 (0.5) | 0.05 | 0.01 |
| Previous outpatient mental health-related service use, n (%) | 229 678 (42.4) | 292 740 (25.4) | 169 618 (20.2) | 0.37 | 0.12 |
| Paediatric complex chronic condition, n (%) | 40 183 (7.4) | 58 343 (5.1) | 36 836 (4.4) | 0.10 | 0.03 |
| Visit with rostered physician, n (%) | |||||
| Yes | 239 875 (44.3) | 687 003 (59.6) | 306 557 (36.6) | 0.31 | 0.47 |
| No | 301 614 (55.7) | 464 871 (40.4) | 531 472 (63.4) | 0.31 | 0.47 |
| Median (IQR) number of outpatient visits for each child during the study | 4 (2–7) | 4 (2–7) | 4 (3–7) | 0.03 | 0.10 |
| Mean±SD continuity of care index‡ | 0.45±0.34§ | 0.44±0.35¶ | 0.33±0.32** | 0.02 | 0.31 |
Antibiotic initiation rate was defined as the sum of each physician’s antibiotic prescriptions for children aged 6–12 years during the study period divided by the total number of visits for children aged 6–12 years during the study period.
Mean standardised differences≥0.10 indicate potentially clinically important differences between groups and are shown in bold.
Calculated for children with>2 outpatient visits during the study period.
N=496 745.
N=1 045 841.
N=772 733.
n, number.
Characteristics of physicians and children by physician psychotropic drug prescribing intensity
There were no differences in the distribution of physicians’ sex, age or years practising medicine by psychotropic drug prescribing intensity group (table 3). High-intensity psychotropic drug prescribers had a higher proportion of Canadian medical school graduates, paediatricians and physicians predominantly remunerated under salaried or capitation-based models, and they had lower visit volumes for 6–12-year-olds compared with average-intensity prescribers.
Table 3. Characteristics of physicians by their psychotropic drug prescribing intensity, 2018–2019, Ontario.
| Physician characteristic | Physician psychotropic drug prescribing intensity | Mean standardised difference* | |||
|---|---|---|---|---|---|
| Low N=2572 |
Average N=5142 |
High N=2572 |
Low versus average | High versus average | |
| Sex, n (%) | |||||
| Female | 1258 (48.9) | 2545 (49.5) | 1291 (50.2) | 0.01 | 0.01 |
| Male | 1314 (51.1) | 2597 (50.5) | 1281 (49.8) | 0.01 | 0.01 |
| Age group, n (%) | |||||
| <45 years | 514 (20.0) | 1128 (21.9) | 579 (22.5) | 0.05 | 0.01 |
| 45–64 years | 1128 (43.9) | 2360 (45.9) | 1226 (47.7) | 0.04 | 0.04 |
| ≥65 years | 390 (15.2) | 604 (11.7) | 305 (11.9) | 0.10 | 0 |
| Missing | 540 (21.0) | 1050 (20.4) | 462 (18.0) | 0.01 | 0.06 |
| Years practising, n (%) | |||||
| ≤5 years | 269 (10.5) | 542 (10.5) | 254 (9.9) | 0 | 0.02 |
| 6–10 years | 370 (14.4) | 707 (13.7) | 413 (16.1) | 0.02 | 0.06 |
| 11–25 years | 678 (26.4) | 1677 (32.6) | 768 (29.9) | 0.14 | 0.06 |
| >25 years | 1243 (48.3) | 2189 (42.6) | 1123 (43.7) | 0.12 | 0.02 |
| Missing | 12 (0.5) | 27 (0.5) | 14 (0.5) | 0.01 | 0 |
| Main speciality, n (%) | |||||
| Family medicine | 2482 (96.5) | 4832 (94.0) | 2105 (81.8) | 0.12 | 0.38 |
| Paediatrics | 90 (3.5) | 310 (6.0) | 467 (18.2) | 0.12 | 0.38 |
| Country of medical training, n (%) | |||||
| Canada | 1114 (43.3) | 2744 (53.4) | 1648 (64.1) | 0.20 | 0.22 |
| USA | 18 (0.7) | 33 (0.6) | 15 (0.6) | 0.01 | 0.01 |
| International | 900 (35.0) | 1315 (25.6) | 447 (17.4) | 0.21 | 0.20 |
| Missing | 540 (21.0) | 1050 (20.4) | 462 (18.0) | 0.01 | 0.06 |
| Predominantly fee-for-service remuneration model, n (%) | |||||
| Yes | 1506 (58.6) | 2297 (44.7) | 860 (33.4) | 0.28 | 0.23 |
| No | 1066 (41.4) | 2845 (55.3) | 1712 (66.6) | 0.28 | 0.23 |
| Rural practice setting, n (%) | 108 (4.2) | 362 (7.0) | 312 (12.1) | 0.12 | 0.17 |
| Median (IQR) number of visits for 6–12-year-olds during the study | 113 (53–306) | 137 (76–270) | 104 (60–181) | 0.15 | 0.28 |
| Mean±SD per cent of visits for 6–12-year-olds during the study | 4.3±4.5 | 5.3±6.0 | 8.7±12.9 | 0.18 | 0.34 |
Mean standardised differences≥0.10 indicate potentially clinically important differences between groups and are shown in bold.
n, number.
High-intensity psychotropic drug prescribers (vs average-intensity prescribers) cared for a higher proportion of male children, Canadian-born children (vs immigrants/refugees), children with previous mental health-related outpatient service use, children with a higher continuity of care index and children living in rural areas or areas with higher incomes or the lowest marginalisation on the racial and newcomer populations dimensions (table 4). The characteristics of physicians and children by prescribing intensity group for each psychotropic drug class are shown in online supplemental tables S1 and S2.
Table 4. Characteristics of child visits by physician psychotropic drug prescribing intensity, 2018–2019, Ontario.
| Child characteristics | Physician psychotropic drug prescribing intensity | Mean standardised difference* | |||
|---|---|---|---|---|---|
| Low N=6 50 279 |
Average N=1 319 259 |
High N=5 61 854 |
Low versus average | High versus average | |
| Sex, n (%) | |||||
| Female | 316 612 (48.7) | 652 658 (49.5) | 245 565 (43.7) | 0.02 | 0.12 |
| Male | 333 667 (51.3) | 666 601 (50.5) | 316 289 (56.3) | 0.02 | 0.12 |
| Mean±SD age, years | 8.9±2.0 | 8.8±2.0 | 8.9±2.0 | 0 | 0.03 |
| Rural residence, n (%) | 7717 (1.2) | 53 087 (4.0) | 59 616 (10.6) | 0.18 | 0.25 |
| Immigration status, n (%) | |||||
| Born in Canada | 562 635 (86.5) | 1 222 223 (92.6) | 545 643 (97.1) | 0.20 | 0.20 |
| Immigrant | 66 844 (10.3) | 70 161 (5.3) | 11 405 (2.0) | 0.19 | 0.18 |
| Refugee | 20 800 (3.2) | 26 875 (2.0) | 4806 (0.9) | 0.07 | 0.10 |
| Income quintile, n (%) | |||||
| 1 (lowest) | 146 473 (22.5) | 217 757 (16.5) | 95 073 (16.9) | 0.15 | 0.01 |
| 2 | 134 824 (20.7) | 213 741 (16.2) | 97 469 (17.3) | 0.12 | 0.03 |
| 3 | 154 160 (23.7) | 258 181 (19.6) | 108 848 (19.4) | 0.10 | 0 |
| 4 | 126 975 (19.5) | 309 859 (23.5) | 126 706 (22.6) | 0.10 | 0.02 |
| 5 (highest) | 86 330 (13.3) | 316 474 (24.0) | 132 115 (23.5) | 0.28 | 0.01 |
| Missing | 1517 (0.2) | 3247 (0.2) | 1643 (0.3) | 0 | 0.01 |
| Households and dwellings quintile, n (%) | |||||
| 1 (least marginalised) | 257 157 (39.5) | 418 778 (31.7) | 127 779 (22.7) | 0.16 | 0.20 |
| 2 | 103 583 (15.9) | 258 678 (19.6) | 121 483 (21.6) | 0.10 | 0.05 |
| 3 | 83 267 (12.8) | 221 317 (16.8) | 110 828 (19.7) | 0.11 | 0.08 |
| 4 | 84 418 (13.0) | 197 725 (15.0) | 110 573 (19.7) | 0.06 | 0.12 |
| 5 (most marginalised) | 119 738 (18.4) | 216 239 (16.4) | 86 330 (15.4) | 0.05 | 0.03 |
| Missing | 2116 (0.3) | 6522 (0.5) | 4861 (0.9) | 0.03 | 0.05 |
| Material resources quintile, n (%) | |||||
| 1 (least marginalised) | 100 213 (15.4) | 345 080 (26.2) | 149 207 (26.6) | 0.27 | 0.01 |
| 2 | 130 996 (20.1) | 310 385 (23.5) | 120 044 (21.4) | 0.08 | 0.05 |
| 3 | 143 608 (22.1) | 234 895 (17.8) | 95 615 (17.0) | 0.11 | 0.02 |
| 4 | 130 343 (20.0) | 200 164 (15.2) | 87 088 (15.5) | 0.13 | 0.01 |
| 5 (most marginalised) | 143 003 (22.0) | 222 213 (16.8) | 105 039 (18.7) | 0.13 | 0.05 |
| Missing | 2116 (0.3) | 6522 (0.5) | 4861 (0.9) | 0.03 | 0.05 |
| Racial and newcomer populations quintile, n (%) | |||||
| 1 (least marginalised) | 10 714 (1.6) | 91 642 (6.9) | 95 119 (16.9) | 0.26 | 0.31 |
| 2 | 20 299 (3.1) | 149 260 (11.3) | 114 393 (20.4) | 0.32 | 0.25 |
| 3 | 45 218 (7.0) | 224 692 (17.0) | 120 520 (21.5) | 0.31 | 0.11 |
| 4 | 107 796 (16.6) | 339 255 (25.7) | 124 753 (22.2) | 0.23 | 0.08 |
| 5 (most marginalised) | 464 136 (71.4) | 507 888 (38.5) | 102 208 (18.2) | 0.70 | 0.46 |
| Missing | 2116 (0.3) | 6522 (0.5) | 4861 (0.9) | 0.03 | 0.05 |
| Age and labour force quintile, n (%) | |||||
| 1 (least marginalised) | 297 056 (45.7) | 522 551 (39.6) | 168 524 (30.0) | 0.12 | 0.20 |
| 2 | 149 458 (23.0) | 285 713 (21.7) | 109 315 (19.5) | 0.03 | 0.05 |
| 3 | 87 224 (13.4) | 205 738 (15.6) | 98 733 (17.6) | 0.06 | 0.05 |
| 4 | 71 210 (11.0) | 172 303 (13.1) | 93 481 (16.6) | 0.06 | 0.10 |
| 5 (most marginalised) | 43 215 (6.6) | 126 432 (9.6) | 86 940 (15.5) | 0.11 | 0.18 |
| Missing | 2116 (0.3) | 6522 (0.5) | 4861 (0.9) | 0.03 | 0.05 |
| Previous outpatient mental health-related service use, n (%) | 104 669 (16.1) | 311 070 (23.6) | 276 297 (49.2) | 0.19 | 0.55 |
| Paediatric complex chronic condition, n (%) | 25 498 (3.9) | 68 185 (5.2) | 41 679 (7.4) | 0.06 | 0.09 |
| Visit with rostered physician, n (%) | |||||
| Yes | 343 069 (52.8) | 626 357 (47.5) | 264 009 (47.0) | 0.11 | 0.01 |
| No | 307 210 (47.2) | 692 902 (52.5) | 297 845 (53.0) | 0.11 | 0.01 |
| Median (IQR) number of outpatient visits for each child during the study | 5 (3–8) | 4 (2–7) | 4 (2–7) | 0.18 | 0.02 |
| Mean±SD continuity of care index† | 0.42±0.34‡ | 0.37±0.34§ | 0.47±0.35¶ | 0.13 | 0.29 |
Mean standardised differences≥0.10 indicate potentially clinically important differences between groups and are shown in bold.
Calculated for children with >2 outpatient visits during the study period.
N=608 827.
N=1 197 923.
N=508 569.
n, number.
Correlation between antibiotic and psychotropic drug prescribing rates
14% of high-intensity antibiotic prescribers were also high-intensity prescribers of study psychotropic drugs; 14%, 21% and 25% were high-intensity prescribers of stimulants, antidepressants and antipsychotics, respectively. Physicians’ antibiotic initiation rate was weakly negatively correlated with their study psychotropic drug prescribing rate (Spearman’s ρ=−0.14). Corresponding values for correlations between physicians’ antibiotic initiation rate and their stimulant, antidepressant and antipsychotic prescribing rates were −0.13 to −0.003 and 0.02, respectively. Antibiotic and psychotropic drug prescribing quartiles are shown in online supplemental table S3.
Association between high-intensity antibiotic prescribing and prescribing of psychotropic drugs
The odds of a physician prescribing any study psychotropic drug, a stimulant or an antidepressant were lower among high-intensity antibiotic prescribers compared with average/low-intensity prescribers (crude OR 0.65 (95% CI 0.60 to 0.70), 0.66 (95% CI 0.61 to 0.72) and 0.75 (95% CI 0.66 to 0.85), respectively). This inverse association persisted after adjusting for child characteristics (adjusted OR (aOR) 0.66 (95% CI 0.62 to 0.71), 0.67 (95% CI 0.62 to 0.72) and aOR 0.79 (0.70 to 0.89), respectively) (table 5). There was no association between the odds of a physician prescribing an antipsychotic and high-intensity antibiotic prescribing (aOR 0.95 (95% CI 0.84 to 1.08)).
Table 5. Association between high-intensity antibiotic prescribing* by physicians and their likelihood of prescribing psychotropic drugs to children, 2018–2019, Ontario.
| Stimulants, antidepressants and/or antipsychotics | Stimulants | Antidepressants | Antipsychotics | |||||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | aOR (95% CI) | OR (95% CI) | aOR (95% CI) | OR (95% CI) | aOR (95% CI) | OR (95% CI) | aOR (95% CI) | |
| Primary analysis population | 0.65 (0.60 to 0.70) | 0.66 (0.62 to 0.71) | 0.66 (0.61 to 0.72) | 0.67 (0.62 to 0.72) | 0.75 (0.66 to 0.85) | 0.79 (0.70 to 0.89) | 0.97 (0.85 to 1.11) | 0.95 (0.84 to 1.08) |
| Continuity of care index | ||||||||
| High† | 0.41 (0.33 to 0.53) | 0.42 (0.34 to 0.52) | 0.41 (0.31 to 0.53) | 0.39 (0.31 to 0.49) | 0.65 (0.43 to 1.00) | 0.68 (0.45 to 1.02) | 0.76 (0.47 to 1.23) | 0.75 (0.47 to 1.20) |
| Average/low‡ | 0.71 (0.65 to 0.77) | 0.72 (0.66 to 0.78) | 0.73 (0.66 to 0.80) | 0.73 (0.67 to 0.80) | 0.79 (0.68 to 0.91) | 0.80 (0.70 to 0.92) | 0.91 (0.78 to 1.07) | 0.90 (0.78 to 1.03) |
| Visit with rostered primary care physician | ||||||||
| Yes | 0.53 (0.48 to 0.59) | 0.54 (0.48 to 0.59) | 0.52 (0.47 to 0.59) | 0.53 (0.48 to 0.58) | 0.70 (0.58 to 0.84) | 0.72 (0.59 to 0.86) | 0.76 (0.60 to 0.98) | 0.74 (0.58 to 0.94) |
| No | 0.75 (0.68 to 0.83) | 0.74 (0.68 to 0.81) | 0.75 (0.68 to 0.84) | 0.74 (0.67 to 0.81) | 0.76 (0.65 to 0.89) | 0.75 (0.65 to 0.88) | 0.85 (0.72 to 1.01) | 0.84 (0.71 to 0.98) |
| Primarily fee-for-service remuneration model | ||||||||
| Yes | 0.63 (0.56 to 0.71) | 0.63 (0.56 to 0.70) | 0.63 (0.56 to 0.71) | 0.63 (0.56 to 0.70) | 0.71 (0.59 to 0.85) | 0.71 (0.59 to 0.85) | 0.92 (0.76 to 1.13) | 0.91 (0.75 to 1.10) |
| No | 0.86 (0.76 to 0.97) | 0.86 (0.77 to 0.97) | 0.87 (0.76 to 0.99) | 0.87 (0.77 to 0.98) | 1.02 (0.83 to 1.25) | 1.03 (0.84 to 1.27) | 1.19 (0.94 to 1.51) | 1.15 (0.92 to 1.44) |
| Exclusion of outlier antibiotic prescribers§ | 0.69 (0.64 to 0.75) | 0.69 (0.64 to 0.75) | 0.70 (0.65 to 0.76) | 0.70 (0.65 to 0.76) | 0.76 (0.67 to 0.87) | 0.77 (0.68 to 0.88) | 1.01 (0.88 to 1.17) | 0.99 (0.86 to 1.13) |
Antibiotic initiation rate was defined as the sum of each physician’s antibiotic prescriptions for children aged 6–12 years during the study period divided by their total number of visits for children aged 6–12 years during the study period.
Quartile 4.
Quartiles 1–3.
Physicians falling below the fifth percentile for antibiotic initiations were excluded.
aOR, adjusted OR.
Exploratory analyses
The inverse association between the prescribing of study psychotropic drugs and high-intensity antibiotic prescribing was generally more pronounced for visits with children with a high versus an average/low continuity of care index (figure 2, table 5). Likewise, the inverse association was stronger for visits with children’s rostered primary care physicians and for visits with physicians remunerated under a fee-for-service model compared with salaried or capitation-based models. The exclusion of outlier antibiotic prescribers had minimal impact on results.
Figure 2. Subgroup analyses of the association between physician high-intensity antibiotic prescribing and their likelihood of prescribing a study psychotropic drug among children, 2018–2019, Ontario. *Quartile 4. †Quartiles 1–3. ‡Physicians falling below the fifth percentile for antibiotic initiations were excluded.
Discussion
We leveraged a large population-based repository of linked clinical, sociodemographic and prescription data to examine the association between physicians’ antibiotic prescribing intensity and their propensity to prescribe psychotropic drugs to children. Previous studies in adults have found that physicians’ prescribing behaviour extends across multiple classes of medications and is largely independent of patient characteristics.19,22 In our analyses of more than 2.5 million primary care visits for children aged 6–12 years in Ontario, we found that physicians who prescribed high volumes of antibiotics were generally less likely to prescribe study psychotropic drugs, with some variability in the strength of the association by drug class (stimulants>antidepressants>>antipsychotics). Results were mostly unchanged after accounting for a wide range of child characteristics.
Several observations might explain these findings. The inverse association between high-intensity antibiotic prescribing and the odds of prescribing a psychotropic drug was stronger when the analysis was restricted to visits with children’s rostered primary care physicians or children with higher care continuity. This might suggest that when a child with psychological or behavioural problems is unknown and follow-up is uncertain, physicians are reluctant to prescribe a long-term medication that requires knowledge of the child’s medical and mental health history, contextual information from parents and other caregivers and ongoing monitoring. This careful, individualised approach to paediatric prescribing may partly explain why prescribing patterns for children differ from those for adults. In adults, high antibiotic prescribing was often correlated with high prescribing of other long-term medications,19,21 a pattern that did not hold true for children in this study. By contrast, when a child with symptoms of an infection is unknown and the ability to follow up is uncertain, physicians might be more likely to prescribe an antibiotic ‘just in case’. One US study found that physicians prescribed more antibiotics when the patient was not theirs34 and another found that patients seen at urgent care centres, where physicians and patients typically have no prior relationship, were more likely to receive antibiotics unnecessarily compared with patients treated for the same illness at traditional ambulatory care clinics.35 By contrast, a Norwegian study found that physicians were three times more likely to use expectant management for a variety of medical problems when they rated their knowledge of their patient as good or excellent versus scant.36 Like our study, the findings from the Norwegian study also suggest that physicians’ accumulated knowledge of their patients influences prescribing behaviour in both directions. For example, among encounters in the Norwegian study for which a prescription was considered, when physicians had good or excellent knowledge of their patient, their knowledge influenced them to prescribe more restrictively in 58% of encounters but more liberally in 42%.36 When physicians lacked previous knowledge of their patient, their prescribing was more restrictive in 75% of cases, and they were more likely to prescribe restrictively for psychological problems compared with ear or skin infections.36
We also found that the inverse association between high antibiotic prescribing and psychotropic drug prescribing was more pronounced among physicians remunerated under a fee-for-service model compared with salaried or capitation-based models. Physician reimbursement can drive practice volumes, with fee-for-service models incentivising seeing more patients and shorter appointments.37 Multiple studies have reported that high patient volumes are associated with elevated rates of antibiotic prescribing.38 39 In a busy clinic, it might be quicker for a physician to prescribe an antibiotic than explain to a parent why antibiotics are not needed for their child’s symptoms. Physicians remunerated under capitation-based models, by contrast, are incentivised to deliver more comprehensive care and preventative interventions, such as immunisation, which can decrease the need for antibiotics.37 Moreover, physicians remunerated under capitation-based models in Ontario often work within collaborative, interdisciplinary, team-based practices,37 which might be more conducive to managing behavioural and psychiatric illnesses in children.
Many characteristics of high antibiotic prescribers and high psychotropic drug prescribers were different, with little overlap between the two groups. For example, consistent with previous findings,19 a higher proportion of high-intensity antibiotic prescribers were international medical school graduates. By contrast, a higher proportion of high-intensity psychotropic drug prescribers graduated from a Canadian medical school, illustrating that the influence of training environment on prescribing behaviours is not uniform across drug classes. We also found that a higher proportion of high-intensity antibiotic prescribers were certified in family medicine, whereas high psychotropic drug prescribers were more often paediatricians. Higher antibiotic prescribing for children by family physicians compared with paediatricians has been reported by others40 and might be due to the former seeing fewer paediatric patients and having less confidence in their diagnostic skills for children.41 Higher psychotropic drug prescribing among paediatricians, by contrast, might suggest children with complex psychiatric or behavioural problems are commonly referred to paediatricians for assessment and treatment. Like several previous studies,19 20 we found that male sex, older age and longer time since medical school graduation were more common among high antibiotic prescribers. By contrast, the distribution of these characteristics did not differ by physician psychotropic drug prescribing intensity. Others have also found that the propensity of physicians to prescribe psychotropic drugs was not predicted by sex or age.42
Sociodemographic characteristics of children were mostly comparable across different levels of their physicians’ antibiotic prescribing intensity, whereas there were multiple differences in the sociodemographic characteristics of children cared for by low-intensity, average-intensity and high-intensity psychotropic drug prescribers. High psychotropic drug prescribers cared for a higher proportion of Canadian-born children versus immigrant or refugee children. Likewise, high-intensity psychotropic drug prescribers cared for a higher proportion of children living in areas with the lowest marginalisation on the racial and newcomer populations dimension. A higher reported burden of mental illness among Canadian-born children compared with immigrant and refugee children has been reported in multiple studies and could reflect differences in healthcare access or health-seeking behaviour, cultural differences in conceptualisation of mental illness or selective migration of healthier families (ie, ‘healthy immigrant effect’).43 44 High-intensity psychotropic drug prescribers also cared for more children residing in rural areas, which could reflect the limited availability of non-pharmacological therapies in rural and remote regions of Canada and difficulties recruiting mental healthcare professionals in these regions.45
Our study has certain limitations. Our measure of antibiotic prescribing intensity is highly correlated with unnecessary antibiotic prescribing;30 however, it is unclear whether the higher propensity of some physicians in this study to prescribe psychotropic drugs for children should raise concerns about overprescribing, or whether the opposite tendency (ie, lower prescribing rates) might signal undertreatment of mental health conditions, or a combination of these and other factors. Relatedly, there is no universally agreed definition of ‘high-intensity prescribing’. We used the top quartile because it has been validated against other standard metrics of antibiotic prescribing.31 Although we adjusted for multiple area-level indicators of marginalisation, we lacked information on the regional availability of community support networks, school-based programmes and other non-pharmacological treatments. Similarly, we did not have access to individual-level data on certain characteristics that have been linked with psychotropic drug use among children, such as foster-care placement46 and juvenile justice system involvement.47 Primary care data from community health centres, which typically serve a higher proportion of marginalised populations, were not available, although community health centres typically serve <1% of the Ontario population.43 We could not ascertain if psychotropic drugs were initiated by psychiatrists or other mental health specialists who were not included in this study. In Ontario, attention deficit/hyperactivity disorder diagnoses are primarily made by family physicians and paediatricians (75%),48 although specialists might have a larger role in the management of other psychological or behavioural problems in young children. This might partially explain the weaker negative association between high-intensity antibiotic prescribing and antidepressant or antipsychotic prescribing compared with stimulants. We could not adjust our statistical models for previous mental health service use since it was highly collinear with receipt of a psychotropic drug prescription. Finally, our data were from a single Canadian province during a period of universal pharmacare for children, potentially limiting generalisability.
Conclusions
We found that physicians who prescribed high volumes of antibiotics for children were generally less likely to prescribe study psychotropic drugs. Factors that shape physician prescribing behaviour are complex, and our results suggest that factors such as care continuity and reimbursement model can influence prescribing behaviour in both directions. Our data do not support a ‘one size fits all’ approach to paediatric drug stewardship across drug classes. Instead, further research is needed to inform tailored approaches to address both overprescribing and underprescribing among children.
Supplementary material
The analyses, conclusions, opinions and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred. Adapted from Statistics Canada, Canadian Census 2016. This does not constitute an endorsement by Statistics Canada of this product. We thank the Toronto Community Health Profiles Partnership for providing access to the Ontario Marginalisation Index.
Footnotes
Funding: This work was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long-Term Care (grant number n/a). This study also received funding from the Canadian Institutes of Health Research (grant number n/a). SCJJ was supported by a grant from the Canadian Immunisation Research Network (grant number n/a).
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Ethics approval: ICES is a prescribed entity under Ontario’s Personal Health Information Protection Act (PHIPA). Section 45 of PHIPA authorises ICES to collect personal health information, without consent, for the purpose of analysis or compiling statistical information with respect to the management of, evaluation of or monitoring of the allocation of resources to or planning for all or part of the health system. Projects that use data collected by ICES under section 45 of PHIPA and use no other data are exempt from research ethics board review. The use of the data in this project is authorised under section 45 and approved by ICES’ Privacy and Legal Office.
Patient and public involvement: Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
Data availability free text: The dataset from this study is held securely in coded form at ICES. While legal data-sharing agreements between ICES and data providers (eg, healthcare organisations and government) prohibit ICES from making the dataset publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at https://www.ices.on.ca/DAS (email: das@ices.on.ca). SCJJ had full access to all data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Code availability: the full dataset creation plan and underlying analytic code are available from the authors upon request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification. Correspondence and requests for materials should be addressed to SCJJ.
Data availability statement
Data may be obtained from a third party and are not publicly available.
References
- 1.Morkem R, Patten S, Queenan J, et al. Recent Trends in the Prescribing of ADHD Medications in Canadian Primary Care. J Atten Disord. 2020;24:301–8. doi: 10.1177/1087054717720719. [DOI] [PubMed] [Google Scholar]
- 2.Arora N, Knowles S, Gomes T, et al. Interprovincial Variation in Antipsychotic and Antidepressant Prescriptions Dispensed in the Canadian Pediatric Population. Can J Psychiatry. 2016;61:758–65. doi: 10.1177/0706743716649190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Raman SR, Man KKC, Bahmanyar S, et al. Trends in attention-deficit hyperactivity disorder medication use: a retrospective observational study using population-based databases. Lancet Psychiatry. 2018;5:824–35. doi: 10.1016/S2215-0366(18)30293-1. [DOI] [PubMed] [Google Scholar]
- 4.Olfson M, Blanco C, Liu S-M, et al. National trends in the office-based treatment of children, adolescents, and adults with antipsychotics. Arch Gen Psychiatry. 2012;69:1247–56. doi: 10.1001/archgenpsychiatry.2012.647. [DOI] [PubMed] [Google Scholar]
- 5.Bachmann CJ, Aagaard L, Burcu M, et al. Trends and patterns of antidepressant use in children and adolescents from five western countries, 2005-2012. Eur Neuropsychopharmacol. 2016;26:411–9. doi: 10.1016/j.euroneuro.2016.02.001. [DOI] [PubMed] [Google Scholar]
- 6.Whittington CJ, Kendall T, Fonagy P, et al. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. The Lancet. 2004;363:1341–5. doi: 10.1016/S0140-6736(04)16043-1. [DOI] [PubMed] [Google Scholar]
- 7.Vitiello B, Correll C, van Zwieten-Boot B, et al. Antipsychotics in children and adolescents: increasing use, evidence for efficacy and safety concerns. Eur Neuropsychopharmacol. 2009;19:629–35. doi: 10.1016/j.euroneuro.2009.04.008. [DOI] [PubMed] [Google Scholar]
- 8.National Institute for Health and Care Excellence Attention deficit hyperactivity disorder (update); [d] evidence review for safety of pharmacological treatment; nice guideline cg72; intervention evidence review. 2018. https://www.nice.org.uk/guidance/ng87/evidence/d-pharmacological-safety-pdf-4783686304 Available. [PubMed]
- 9.Kazda L, Bell K, Thomas R, et al. Overdiagnosis of Attention-Deficit/Hyperactivity Disorder in Children and Adolescents: A Systematic Scoping Review. JAMA Netw Open. 2021;4:e215335. doi: 10.1001/jamanetworkopen.2021.5335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Taylor E. Pediatric psychopharmacology: Too much and too little. World Psychiatry. 2013;12:124–5. doi: 10.1002/wps.20030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Schwartz KL, Langford BJ, Daneman N, et al. Unnecessary antibiotic prescribing in a Canadian primary care setting: a descriptive analysis using routinely collected electronic medical record data. CMAJ Open. 2020;8:E360–9. doi: 10.9778/cmajo.20190175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Saatchi A, Marra F. Increasing paediatric prescribing rates in British Columbian children: cause for concern? Can J Public Health. 2020;111:523–30. doi: 10.17269/s41997-020-00298-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Public Health Agency of Canada . Ottawa: Canadian antimicrobial resistance surveillance system: 2017 report.www.canada.ca/en/public-health/services/publications/drugs-health-products/canadian-antimicrobial-resistance-surveillance-system-2017-report-executive-summary.html Available. [Google Scholar]
- 14.Butler AM, Brown DS, Durkin MJ, et al. Association of Inappropriate Outpatient Pediatric Antibiotic Prescriptions With Adverse Drug Events and Health Care Expenditures. JAMA Netw Open. 2022;5:e2214153. doi: 10.1001/jamanetworkopen.2022.14153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Downes KJ, Hayes M, Fitzgerald JC, et al. Mechanisms of antimicrobial-induced nephrotoxicity in children. J Antimicrob Chemother. 2020;75:1–13. doi: 10.1093/jac/dkz325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kim J, Shaklee JF, Smathers S, et al. Risk factors and outcomes associated with severe clostridium difficile infection in children. Pediatr Infect Dis J. 2012;31:134–8. doi: 10.1097/INF.0b013e3182352e2c. [DOI] [PubMed] [Google Scholar]
- 17.Mayne SL, Ross ME, Song L, et al. Variations in Mental Health Diagnosis and Prescribing Across Pediatric Primary Care Practices. Pediatrics. 2016;137:e20152974. doi: 10.1542/peds.2015-2974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Gerber JS, Prasad PA, Russell Localio A, et al. Variation in Antibiotic Prescribing Across a Pediatric Primary Care Network. J Pediatric Infect Dis Soc. 2015;4:297–304. doi: 10.1093/jpids/piu086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Quinn KL, Campitelli MA, Diong C, et al. Association between Physician Intensity of Antibiotic Prescribing and the Prescription of Benzodiazepines, Opioids and Proton-Pump Inhibitors to Nursing Home Residents: a Population-Based Observational Study. J Gen Intern Med. 2019;34:2763–71. doi: 10.1007/s11606-019-05333-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Barrett AK, Cashy JP, Thorpe CT, et al. Latent Class Analysis of Prescribing Behavior of Primary Care Physicians in the Veterans Health Administration. J Gen Intern Med. 2022;37:3346–54. doi: 10.1007/s11606-021-07248-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Li Y, Mölter A, White A, et al. Relationship between prescribing of antibiotics and other medicines in primary care: a cross-sectional study. Br J Gen Pract. 2019;69:e42–51. doi: 10.3399/bjgp18X700457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Evans CT, Fitzpatrick MA, Poggensee L, et al. High prescribing of antibiotics is associated with high prescribing of opioids in medical and dental providers. Pharmacotherapy. 2022;42:716–23. doi: 10.1002/phar.2720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Alessi-Severini S, Biscontri RG, Collins DM, et al. Ten years of antipsychotic prescribing to children: a Canadian population-based study. Can J Psychiatry. 2012;57:52–8. doi: 10.1177/070674371205700109. [DOI] [PubMed] [Google Scholar]
- 24.Ronsley R, Scott D, Warburton WP, et al. A population-based study of antipsychotic prescription trends in children and adolescents in British Columbia, from 1996 to 2011. Can J Psychiatry. 2013;58:361–9. doi: 10.1177/070674371305800608. [DOI] [PubMed] [Google Scholar]
- 25.Levy AR, O’Brien BJ, Sellors C, et al. Coding accuracy of administrative drug claims in the Ontario Drug Benefit database. Can J Clin Pharmacol. 2003;10:67–71. [PubMed] [Google Scholar]
- 26.Matheson FI, Moloney G, VanIngen T. Ontario Marginalization Index: User Guide. Toronto, ON: St. Michael’s Hospital (Unity Health Toronto); 2021. [Google Scholar]
- 27.ICES concept dictionary Morbidity measure - children with medical complexity. :2024.
- 28.Steele LS, Glazier RH, Lin E, et al. Using administrative data to measure ambulatory mental health service provision in primary care. Med Care. 2004;42:960–5. doi: 10.1097/00005650-200410000-00004. [DOI] [PubMed] [Google Scholar]
- 29.Sudhakar-Krishnan V, Rudolf MCJ. How important is continuity of care? Arch Dis Child. 2007;92:381–3. doi: 10.1136/adc.2006.099853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kitano T, Langford BJ, Brown KA, et al. The Association Between High and Unnecessary Antibiotic Prescribing: A Cohort Study Using Family Physician Electronic Medical Records. Clin Infect Dis. 2021;72:e345–51. doi: 10.1093/cid/ciaa1139. [DOI] [PubMed] [Google Scholar]
- 31.Schwartz KL, Chen C, Langford BJ, et al. Validating a popular outpatient antibiotic database to reliably identify high prescribing physicians for patients 65 years of age and older. PLoS ONE. 2019;14:e0223097. doi: 10.1371/journal.pone.0223097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Austin PC. Using the Standardized Difference to Compare the Prevalence of a Binary Variable Between Two Groups in Observational Research. Communications in Statistics - Simulation and Computation. 2009;38:1228–34. doi: 10.1080/03610910902859574. [DOI] [Google Scholar]
- 33.O’Brien KM, Lawrence KG, Keil AP. The Case for Case-Cohort: An Applied Epidemiologist’s Guide to Reframing Case-Cohort Studies to Improve Usability and Flexibility. Epidemiology. 2022;33:354–61. doi: 10.1097/EDE.0000000000001469. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Manne M, Deshpande A, Hu B, et al. Provider Variation in Antibiotic Prescribing and Outcomes of Respiratory Tract Infections. South Med J. 2018;111:235–42. doi: 10.14423/SMJ.0000000000000795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Palms DL, Hicks LA, Bartoces M, et al. Comparison of Antibiotic Prescribing in Retail Clinics, Urgent Care Centers, Emergency Departments, and Traditional Ambulatory Care Settings in the United States. JAMA Intern Med. 2018;178:1267–9. doi: 10.1001/jamainternmed.2018.1632. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Hjortdahl P, Borchgrevink CF. Continuity of care: influence of general practitioners’ knowledge about their patients on use of resources in consultations. BMJ. 1991;303:1181–4. doi: 10.1136/bmj.303.6811.1181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Hong M, Devlin RA, Zaric GS, et al. Primary care services and emergency department visits in blended fee-for-service and blended capitation models: evidence from Ontario, Canada. Eur J Health Econ. 2024;25:363–77. doi: 10.1007/s10198-023-01591-w. [DOI] [PubMed] [Google Scholar]
- 38.Aabenhus R, Siersma V, Sandholdt H, et al. Identifying practice-related factors for high-volume prescribers of antibiotics in Danish general practice. J Antimicrob Chemother. 2017;72:2385–91. doi: 10.1093/jac/dkx115. [DOI] [PubMed] [Google Scholar]
- 39.Gjelstad S, Straand J, Dalen I, et al. Do general practitioners’ consultation rates influence their prescribing patterns of antibiotics for acute respiratory tract infections? J Antimicrob Chemother. 2011;66:2425–33. doi: 10.1093/jac/dkr295. [DOI] [PubMed] [Google Scholar]
- 40.Nyquist AC, Gonzales R, Steiner JF, et al. Antibiotic prescribing for children with colds, upper respiratory tract infections, and bronchitis. JAMA. 1998;279:875–7. doi: 10.1001/jama.279.11.875. [DOI] [PubMed] [Google Scholar]
- 41.Schwartz B, Mainous AG, 3rd, Marcy SM Why Do Physicians Prescribe Antibiotics for Children With Upper Respiratory Tract Infections? JAMA. 1998;279:881. doi: 10.1001/jama.279.11.881. [DOI] [PubMed] [Google Scholar]
- 42.Hadsall RS, Freeman RA, Joseph Norwood G. Factors related to the prescribing of selected psychotropic drugs by primary care physicians. Soc Sci Med. 1982;16:1747–56. doi: 10.1016/0277-9536(82)90268-4. [DOI] [PubMed] [Google Scholar]
- 43.Saunders NR, Lebenbaum M, Lu H, et al. Trends in mental health service utilisation in immigrant youth in Ontario, Canada, 1996-2012: a population-based longitudinal cohort study. BMJ Open. 2018;8:e022647. doi: 10.1136/bmjopen-2018-022647. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Gadermann AM, Gagné Petteni M, Janus M, et al. Prevalence of Mental Health Disorders Among Immigrant, Refugee, and Nonimmigrant Children and Youth in British Columbia, Canada. JAMA Netw Open. 2022;5:e2144934. doi: 10.1001/jamanetworkopen.2021.44934. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Furlan AD, Pajer KA, Gardner W, et al. Project ECHO: Building capacity to manage complex conditions in rural, remote and underserved areas. Can J Rural Med. 2019;24:115–20. doi: 10.4103/CJRM.CJRM_20_18. [DOI] [PubMed] [Google Scholar]
- 46.Chiang Y-Y, Amill-Rosario A, Tran P, et al. Psychotropic Polypharmacy Among Youths Enrolled in Medicaid. JAMA Netw Open. 2024;7:e2356404. doi: 10.1001/jamanetworkopen.2023.56404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Cuellar AE, Kelleher KJ, Kataoka S, et al. Incarceration and Psychotropic Drug Use by Youth. Arch Pediatr Adolesc Med. 2008;162:219. doi: 10.1001/archpediatrics.2007.47. [DOI] [PubMed] [Google Scholar]
- 48.Vasiliadis H-M, Diallo FB, Rochette L, et al. Temporal Trends in the Prevalence and Incidence of Diagnosed ADHD in Children and Young Adults between 1999 and 2012 in Canada: A Data Linkage Study. Can J Psychiatry. 2017;62:818–26. doi: 10.1177/0706743717714468. [DOI] [PMC free article] [PubMed] [Google Scholar]