| Characteristic | Female N = 5,2441 |
Male N = 2,2631 |
Overall N = 7,5071 |
p-value2 |
|---|---|---|---|---|
| Age (years) | 31.0 ± 18.0 | 30.3 ± 18.3 | 30.8 ± 18.1 | 0.376 |
| Highest Educational Qualification | <0.001 | |||
| No formal education | 658 (12.5%) | 174 (7.7%) | 832 (11.1%) | |
| Primary | 371 (7.1%) | 164 (7.2%) | 535 (7.1%) | |
| Junior High School | 435 (8.3%) | 133 (5.9%) | 568 (7.6%) | |
| Middle School Leaving Certificate | 176 (3.4%) | 92 (4.1%) | 268 (3.6%) | |
| Senior High School | 683 (13.0%) | 235 (10.4%) | 918 (12.2%) | |
| Tertiary | 2,921 (55.7%) | 1,465 (64.7%) | 4,386 (58.4%) | |
| Type of Patient | <0.001 | |||
| Corporate | 163 (3.1%) | 65 (2.9%) | 228 (3.0%) | |
| KNUST Staff | 168 (3.2%) | 196 (8.7%) | 364 (4.8%) | |
| KNUST Staff Dependant | 497 (9.5%) | 124 (5.5%) | 621 (8.3%) | |
| KNUST Student | 1,915 (36.5%) | 1,009 (44.6%) | 2,924 (39.0%) | |
| Private | 2,501 (47.7%) | 869 (38.4%) | 3,370 (44.9%) | |
| Insurance Status (NHIS) | <0.001 | |||
| Insured | 4,285 (81.7%) | 1,640 (72.5%) | 5,925 (78.9%) | |
| Not Insured | 959 (18.3%) | 623 (27.5%) | 1,582 (21.1%) | |
| Current NHIS Status | <0.001 | |||
| Active | 3,826 (73.0%) | 1,413 (62.4%) | 5,239 (69.8%) | |
| Inactive | 459 (8.8%) | 227 (10.0%) | 686 (9.1%) | |
| Never Enrolled | 959 (18.3%) | 623 (27.5%) | 1,582 (21.1%) | |
| Has comorbidity | 1,796 (34.2%) | 593 (26.2%) | 2,389 (31.8%) | <0.001 |
| Hypertension | 742 (14.1%) | 244 (10.8%) | 986 (13.1%) | <0.001 |
| Arthritis | 674 (12.9%) | 206 (9.1%) | 880 (11.7%) | <0.001 |
| Peripheral neuropathy | 410 (7.8%) | 108 (4.8%) | 518 (6.9%) | <0.001 |
| Diabetes mellitus | 322 (6.1%) | 76 (3.4%) | 398 (5.3%) | <0.001 |
| Pregnancy | 374 (7.1%) | 1 (0.0%) | 375 (5.0%) | <0.001 |
| Peptic ulcer disease | 166 (3.2%) | 50 (2.2%) | 216 (2.9%) | 0.023 |
| Prostate enlargement | 0 (0.0%) | 131 (5.8%) | 131 (1.7%) | <0.001 |
| Asthma | 72 (1.4%) | 21 (0.9%) | 93 (1.2%) | 0.110 |
| Anaemia | 58 (1.1%) | 14 (0.6%) | 72 (1.0%) | 0.047 |
| Sickle cell disease | 41 (0.8%) | 23 (1.0%) | 64 (0.9%) | 0.311 |
| Stroke | 32 (0.6%) | 19 (0.8%) | 51 (0.7%) | 0.267 |
| Renal calculus | 15 (0.3%) | 22 (1.0%) | 37 (0.5%) | <0.001 |
| Renal colic | 9 (0.2%) | 16 (0.7%) | 25 (0.3%) | <0.001 |
| Chronic kidney disease | 12 (0.2%) | 12 (0.5%) | 24 (0.3%) | 0.034 |
| Hepatitis | 7 (0.1%) | 17 (0.8%) | 24 (0.3%) | <0.001 |
| Heart failure | 11 (0.2%) | 8 (0.4%) | 19 (0.3%) | 0.255 |
| Liver disease | 10 (0.2%) | 7 (0.3%) | 17 (0.2%) | 0.321 |
| Cancer | 11 (0.2%) | 2 (0.1%) | 13 (0.2%) | 0.367 |
| HIV/AIDS | 8 (0.2%) | 4 (0.2%) | 12 (0.2%) | 0.761 |
| Urinary retention | 2 (0.0%) | 8 (0.4%) | 10 (0.1%) | 0.002 |
| Other | 19 (0.4%) | 13 (0.6%) | 32 (0.4%) | 0.195 |
| 1 Mean ± SD; n (%) | ||||
| 2 Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test | ||||
Antibiotic Prescribing Patterns in Urinary Tract Infections
A Retrospective Study at KNUST Hospital, Ghana (January–December 2023)
Note
Key Message UTI antibiotic prescribing in this cohort is badly misaligned with global stewardship targets — and the gap is structural, not just behavioral. The headline numbers, in order of importance:
84% Watch-category vs. an 8% Access-category share (WHO target: ≥70% Access). This is the single most important finding — prescribing is almost the mirror image of where it should be. The gap is baked into local guidelines, not just prescriber choice. Ghana’s own first-line agents for UTI (ciprofloxacin, cefuroxime) are themselves classified as WHO “Watch.” So even a clinician following local guidelines perfectly will still show up as AWaRe-discordant. This reframes the story from “prescribers are doing it wrong” to “the local standard of care and the global framework are in tension” — which is a more useful, less blame-y message for a clinical or policy audience. 81% of antibiotics were started without a culture. Treatment is overwhelmingly empirical, which limits any ability to de-escalate or target therapy later. 39% of courses were too short (vs. only 2.4% too long). The stewardship risk here isn’t overuse — it’s under-treatment, which is a less commonly told but arguably more urgent story for this dataset. A ciprofloxacin/tinidazole combination — explicitly “Not Recommended” — was the 4th most-prescribed antibiotic, ahead of legitimate first-line Access agents. This is a concrete, named, fixable target.
Introduction
Urinary tract infections (UTIs) are among the most prevalent bacterial infections globally, disproportionately affecting women and placing a substantial burden on health systems. In Ghana, as in much of sub-Saharan Africa, the rising prevalence of antimicrobial resistance (AMR) threatens the effectiveness of empirically prescribed antibiotics. Ciprofloxacin and cefuroxime — the first-line oral agents recommended by the Ghana Standard Treatment Guidelines (STG, 7th Edition, 2017) for uncomplicated UTIs — face resistance rates exceeding 60% in Ghanaian clinical isolates, underscoring the urgency of rational prescribing.
Rational antibiotic use is a cornerstone of antimicrobial stewardship. The World Health Organization (WHO) AWaRe (Access, Watch, Reserve) framework classifies antibiotics by their resistance risk profile and recommends that ≥70% of antibiotic consumption in any health system should derive from the Access category — agents with lower resistance potential suitable for common infections. Monitoring prescribing patterns against this target provides a measurable stewardship benchmark.
This study describes antibiotic prescribing patterns for UTIs at KNUST Hospital, Kumasi, Ghana, over a 12-month period in 2023. It evaluates AWaRe target compliance, empirical prescribing rates, treatment duration concordance with the Ghana STG, and patient-level predictors of Watch-category antibiotic prescribing.
Methods
Study setting and design
This was a retrospective review of electronic medical records (EMR) at KNUST Hospital, Kumasi, Ghana, covering all patients diagnosed with a urinary tract infection between January and December 2023.
Data extraction
Records were extracted from the hospital EMR system. Each record included patient demographics (age, sex, education, insurance status), clinical data (diagnosis, department, vital signs, comorbidities), prescription details (antibiotic name, dose, frequency, duration, route), and laboratory investigations requested. Records were linked to institutional price lists for medications and investigations.
Definitions
UTI diagnosis was ascertained from the free-text diagnosis field using pattern matching for UTI-related terms (e.g., “urinary tract infection”, “pyelonephritis”, “cystitis”).
Antibiotic classification followed the WHO AWaRe 2021 classification. Ciprofloxacin/tinidazole combination was manually assigned to the “Not Recommended” category (ATC code J01RA11) as it does not appear in the standard AWaRe list.
Defined Daily Dose (DDD) was calculated using WHO ATC/DDD methodology. Daily dose was derived from the prescribed dose, formulation strength, and frequency. Total dose was computed as daily dose × duration. DDD = total dose / WHO reference DDD. Consumption rates are expressed as DDDs per 1,000 patients per day, with a denominator of 7,507 unique patients over 365 study days.
Empirical prescribing was defined as an antibiotic prescribed at an encounter where no urine culture and sensitivity (C/S) test was requested. Culture-ordered empirical prescribing refers to encounters where a C/S was requested alongside the antibiotic prescription.
Treatment duration concordance was assessed against the Ghana STG 7th Edition (2017) for ciprofloxacin (7 days in females; 10–14 days in males) and cefuroxime (5–7 days in females; 10–14 days in males). IDSA (2011) and WHO Model Formulary (2023) thresholds were applied for agents not specifically addressed by the Ghana STG.
Recurrent UTI was defined as ≥3 episodes in 12 months or ≥2 episodes within 6 months.
Statistical analysis
Categorical variables are presented as frequencies and percentages (n, %). Continuous variables are reported as mean ± SD or median (IQR) depending on distribution. Group comparisons used chi-squared tests, Fisher’s exact test, or the Kruskal-Wallis test as appropriate. A one-sample proportion test evaluated AWaRe Access target compliance against the 70% WHO benchmark. Multivariable logistic regression (presented as odds ratios with 95% confidence intervals) identified independent predictors of Watch or Not-Recommended antibiotic prescribing.
Monthly antibiotic consumption trends were analysed using two time series methods: (i) the Mann-Kendall trend test (non-parametric, tests for monotonic trend; reported as Kendall’s τ with p-value) and (ii) segmented piecewise linear regression (identifies the inflection point and estimates separate slopes for the rising and falling phases, with 95% confidence intervals). Both methods were applied to the overall monthly series and to each AWaRe category separately. All analyses were conducted in R version 4.5.2 using the tidyverse, gtsummary, trend, segmented, and ggplot2 packages.
Results
Patient characteristics
Prescription patterns
The table below summarises antibiotic and non-antibiotic prescribing at the encounter level among UTI patients.
| Characteristic | n (%) N = 288991 |
|---|---|
| Prescription characteristics (drug-line level) | |
| Prescription type | |
| Non-Antibiotic | 19,797 (68.50%) |
| Antibiotic | 8,234 (28.49%) |
| No prescription | 868 (3.00%) |
| Route of administration | |
| Oral | 26,311 (91.04%) |
| Parenteral | 2,334 (8.08%) |
| Topical | 254 (0.88%) |
| Medicines per encounter | |
| Medicines per encounter | 3.0 (2.0, 4.0) |
| Antibiotics per encounter | 1.0 (1.0, 1.0) |
| 1 n (%) | |
Antibiotic use and AWaRe classification
Overall antibiotic distribution
| Variable | N = 8,2391 |
|---|---|
| AWARE Classification | |
| Watch | 6,924 (84.04%) |
| Access | 662 (8.03%) |
| Not Recommended | 653 (7.93%) |
| Class of Antibiotic | |
| Fluoroquinolones | 3,804 (46.17%) |
| Second-generation-cephalosporins | 2,493 (30.26%) |
| Combinations of antibacterials | 653 (7.93%) |
| Third-generation-cephalosporins | 440 (5.34%) |
| Tetracyclines | 358 (4.35%) |
| Macrolides | 181 (2.20%) |
| Beta-lactam/beta-lactamase-inhibitor | 166 (2.01%) |
| Other | 99 (1.20%) |
| Aminoglycosides | 45 (0.55%) |
| Name of Antibiotic | |
| Ciprofloxacin | 3,574 (43.38%) |
| Cefuroxime | 2,493 (30.26%) |
| Ciprofloxacin/tinidazole | 653 (7.93%) |
| Doxycycline | 351 (4.26%) |
| Ceftriaxone | 301 (3.65%) |
| Levofloxacin | 230 (2.79%) |
| Amoxicillin/clavulanic-acid | 166 (2.01%) |
| Azithromycin | 161 (1.95%) |
| Cefixime | 139 (1.69%) |
| Other | 136 (1.65%) |
| Secnidazole | 35 (0.42%) |
| 1 n (%) | |
AWaRe breakdown by category
| Variable |
AWARE Classification
|
||
|---|---|---|---|
| Access N = 6621 |
Watch N = 6,9241 |
Not Recommended N = 6531 |
|
| Class of Antibiotic | |||
| Fluoroquinolones | 0 (0%) | 3,804 (55%) | 0 (0%) |
| Second-generation-cephalosporins | 0 (0%) | 2,493 (36%) | 0 (0%) |
| Combinations of antibacterials | 0 (0%) | 0 (0%) | 653 (100%) |
| Third-generation-cephalosporins | 0 (0%) | 440 (6.4%) | 0 (0%) |
| Tetracyclines | 358 (54%) | 0 (0%) | 0 (0%) |
| Macrolides | 0 (0%) | 181 (2.6%) | 0 (0%) |
| Beta-lactam/beta-lactamase-inhibitor | 166 (25%) | 0 (0%) | 0 (0%) |
| Aminoglycosides | 41 (6.2%) | 4 (<0.1%) | 0 (0%) |
| Imidazoles | 35 (5.3%) | 0 (0%) | 0 (0%) |
| Lincosamides | 29 (4.4%) | 0 (0%) | 0 (0%) |
| Penicillins | 18 (2.7%) | 0 (0%) | 0 (0%) |
| Nitrofuran-derivatives | 12 (1.8%) | 0 (0%) | 0 (0%) |
| Carbapenems | 0 (0%) | 2 (<0.1%) | 0 (0%) |
| Sulfonamide-trimethoprim-combinations | 2 (0.3%) | 0 (0%) | 0 (0%) |
| Amphenicols | 1 (0.2%) | 0 (0%) | 0 (0%) |
| Name of Antibiotic | |||
| Ciprofloxacin | 0 (0%) | 3,574 (52%) | 0 (0%) |
| Cefuroxime | 0 (0%) | 2,493 (36%) | 0 (0%) |
| Ciprofloxacin/tinidazole | 0 (0%) | 0 (0%) | 653 (100%) |
| Doxycycline | 351 (53%) | 0 (0%) | 0 (0%) |
| Ceftriaxone | 0 (0%) | 301 (4.3%) | 0 (0%) |
| Levofloxacin | 0 (0%) | 230 (3.3%) | 0 (0%) |
| Amoxicillin/clavulanic-acid | 166 (25%) | 0 (0%) | 0 (0%) |
| Azithromycin | 0 (0%) | 161 (2.3%) | 0 (0%) |
| Cefixime | 0 (0%) | 139 (2.0%) | 0 (0%) |
| Secnidazole | 35 (5.3%) | 0 (0%) | 0 (0%) |
| Amikacin | 31 (4.7%) | 0 (0%) | 0 (0%) |
| Clindamycin | 29 (4.4%) | 0 (0%) | 0 (0%) |
| Nitrofurantoin | 12 (1.8%) | 0 (0%) | 0 (0%) |
| Amoxicillin | 10 (1.5%) | 0 (0%) | 0 (0%) |
| Clarithromycin | 0 (0%) | 10 (0.1%) | 0 (0%) |
| Erythromycin | 0 (0%) | 10 (0.1%) | 0 (0%) |
| Gentamicin | 10 (1.5%) | 0 (0%) | 0 (0%) |
| Flucloxacillin | 7 (1.1%) | 0 (0%) | 0 (0%) |
| Tetracycline | 7 (1.1%) | 0 (0%) | 0 (0%) |
| Tobramycin | 0 (0%) | 4 (<0.1%) | 0 (0%) |
| Meropenem | 0 (0%) | 2 (<0.1%) | 0 (0%) |
| Sulfamethoxazole/trimethoprim | 2 (0.3%) | 0 (0%) | 0 (0%) |
| Ampicillin | 1 (0.2%) | 0 (0%) | 0 (0%) |
| Chloramphenicol | 1 (0.2%) | 0 (0%) | 0 (0%) |
| 1 n (%) | |||
Antibiotic consumption (Defined Daily Doses)
DDD by AWaRe category, class, and antibiotic
| Variable | N = 8,2391 |
|---|---|
| Aware Classification | |
| Access | 10.00 (4.17, 14.00) |
| Not Recommended | 14.00 (14.00, 14.00) |
| Watch | 7.00 (5.00, 10.00) |
| Antibiotic Class | |
| Aminoglycosides | 1.00 (0.53, 2.50) |
| Amphenicols | 4.67 (4.67, 4.67) |
| Beta-lactam/beta-lactamase-inhibitor | 4.17 (0.01, 5.83) |
| Carbapenems | 0.42 (0.33, 0.50) |
| Combinations of antibacterials | 14.00 (14.00, 14.00) |
| Fluoroquinolones | 7.00 (7.00, 7.00) |
| Imidazoles | 0.00 (0.00, 0.00) |
| Lincosamides | 6.00 (0.67, 7.00) |
| Macrolides | 5.00 (5.00, 5.00) |
| Nitrofuran-derivatives | 7.00 (7.00, 8.50) |
| Penicillins | 7.00 (4.67, 9.33) |
| Second-generation-cephalosporins | 14.00 (7.00, 14.00) |
| Sulfonamide-trimethoprim-combinations | 19.00 (10.00, 28.00) |
| Tetracyclines | 14.00 (14.00, 20.00) |
| Third-generation-cephalosporins | 1.00 (0.50, 6.50) |
| Antibiotic Name | |
| Amikacin | 1.00 (0.50, 2.50) |
| Amoxicillin | 9.33 (5.00, 18.67) |
| Amoxicillin/clavulanic-acid | 4.17 (0.01, 5.83) |
| Ampicillin | 0.30 (0.30, 0.30) |
| Azithromycin | 5.00 (5.00, 5.00) |
| Cefixime | 7.00 (7.00, 7.00) |
| Ceftriaxone | 1.00 (0.25, 1.00) |
| Cefuroxime | 14.00 (7.00, 14.00) |
| Chloramphenicol | 4.67 (4.67, 4.67) |
| Ciprofloxacin | 7.00 (7.00, 7.00) |
| Ciprofloxacin/tinidazole | 14.00 (14.00, 14.00) |
| Clarithromycin | 24.00 (14.00, 28.00) |
| Clindamycin | 6.00 (0.67, 7.00) |
| Doxycycline | 14.00 (14.00, 20.00) |
| Erythromycin | 7.00 (7.00, 14.00) |
| Flucloxacillin | 7.00 (3.50, 7.00) |
| Gentamicin | 1.17 (0.56, 1.33) |
| Levofloxacin | 7.00 (7.00, 10.00) |
| Meropenem | 0.42 (0.33, 0.50) |
| Nitrofurantoin | 7.00 (7.00, 8.50) |
| Secnidazole | 0.00 (0.00, 0.00) |
| Sulfamethoxazole/trimethoprim | 19.00 (10.00, 28.00) |
| Tetracycline | NA (NA, NA) |
| Route of Administration | |
| Oral | 7.00 (7.00, 14.00) |
| Parenteral | 0.50 (0.50, 1.00) |
| 1 Defined Daily Dose (DDD): Median(IQR) | |
Monthly DDD consumption: time series analysis
Monthly antibiotic consumption was analysed using two complementary time series methods: (1) the Mann-Kendall test for monotonic trend and (2) segmented (piecewise linear) regression to identify the inflection point and estimate pre- and post-inflection consumption slopes. Both methods were applied to the overall series and to each AWaRe category separately.
| Time Series Analysis of Monthly Antibiotic Consumption | ||||||
| Segmented regression (piecewise linear) and Mann-Kendall trend test | ||||||
| AWaRe Category |
Segmented Regression
|
Mann-Kendall Test
|
||||
|---|---|---|---|---|---|---|
| Inflection point | Pre-inflection slope1 | Post-inflection slope | R² | Kendall τ | MK p-value | |
| Access | Jun (mo. 6.1) | 0.032 | -0.054 | 0.880 | −0.495 | 0.086 |
| Not Recommended | Jul (mo. 7) | 0.051 | -0.081 | 0.488 | −0.257 | 0.373 |
| Watch | Jun (mo. 6) | 0.263 | -0.297 | 0.828 | −0.297 | 0.304 |
| Total | Jun (mo. 6.2) | 0.343 | -0.432 | 0.817 | −0.337 | 0.244 |
| 1 Slopes in DDDs/1,000 patients/day per month. | ||||||
Segmented regression identified a single inflection point at month 6.2 (June), with an overall model fit of R² = 0.82. Prior to June, antibiotic consumption increased at 0.343 DDDs/1,000 patients/day per month; following June, it declined at 0.432 DDDs/1,000 patients/day per month. The overall Mann-Kendall test did not detect a statistically significant monotonic trend (τ = -0.337, p = 0.244), consistent with the bidirectional (seasonal) rather than unidirectional pattern observed.
Novel analyses
WHO AWaRe Access target compliance
The WHO AWaRe framework recommends that ≥70% of antibiotic prescriptions in any health system derive from the Access category. A one-sample proportion test was used to evaluate whether the observed proportion of Access-category prescriptions met this target.
Access-category antibiotics accounted for only 8% (95% CI: 7.5%–8.6%) of all antibiotic prescriptions — far below the 70% WHO target (one-sample proportion test p < 0.001). Watch-category antibiotics dominated prescribing at 84%, with Not-Recommended agents comprising 7.9%.
Empirical antibiotic prescribing rate
Among 7254 antibiotic encounters, 81% were purely empirical — no urine culture was requested alongside the antibiotic prescription. Only 19% of encounters had a culture ordered concurrently.
| AWaRe Category by Culture-Ordering Status | ||
| AWaRe Category | n | % |
|---|---|---|
| No Culture | ||
| Access | 505 | 7.6 |
| Not Recommended | 470 | 7.1 |
| Watch | 5656 | 85.3 |
| Culture Ordered | ||
| Access | 157 | 9.8 |
| Not Recommended | 183 | 11.4 |
| Watch | 1268 | 78.9 |
Treatment duration concordance with Ghana STG
Prescribed antibiotic duration was compared against the Ghana Standard Treatment Guidelines (7th Ed., 2017) for the two first-line recommended agents (ciprofloxacin and cefuroxime), and against IDSA/WHO thresholds for all other antibiotics. For the Ghana STG agents, thresholds are sex-specific.
Overall, 58.7% of prescriptions were guideline-concordant, 39% were shorter than recommended (under-duration), and 2.4% exceeded the recommended duration.
Predictors of Watch-category antibiotic prescribing
Logistic regression identified patient-level predictors of receiving a Watch or Not-Recommended antibiotic (vs Access category only). The outcome was assigned at the patient level using the highest-concern AWaRe category received.
| Characteristic |
Crude OR (95% CI)
|
Adjusted OR (95% CI)
|
|||||
|---|---|---|---|---|---|---|---|
| N | OR | 95% CI | p-value | OR | 95% CI | p-value | |
| Age (years) | 6,273 | 1.00 | 1.0, 1.01 | 0.482 | 1.01 | 1.00, 1.02 | 0.233 |
| Sex | 6,273 | ||||||
| Female | — | — | — | — | |||
| Male | 0.55 | 0.40, 0.76 | <0.001 | 0.53 | 0.38, 0.73 | <0.001 | |
| Patient Type | 6,273 | ||||||
| Corporate | — | — | — | — | |||
| KNUST Staff | 1.01 | 0.36, 2.68 | 0.981 | 1.09 | 0.38, 2.91 | 0.871 | |
| KNUST Staff Dependant | 0.83 | 0.33, 1.89 | 0.682 | 0.84 | 0.32, 1.93 | 0.697 | |
| KNUST Student | 1.71 | 0.70, 3.57 | 0.192 | 1.83 | 0.73, 3.94 | 0.152 | |
| Private | 1.44 | 0.59, 2.96 | 0.369 | 1.45 | 0.60, 2.99 | 0.364 | |
| Has Comorbidity | 6,273 | ||||||
| No | — | — | — | — | |||
| Yes | 1.03 | 0.74, 1.45 | 0.871 | 0.99 | 0.66, 1.50 | 0.960 | |
| Hypertension | 2,010 | 1.02 | 0.58, 1.83 | 0.938 | |||
| Diabetes Mellitus | 2,010 | 1.80 | 0.78, 5.23 | 0.215 | |||
| Pregnancy | 2,010 | 2.00 | 0.80, 6.66 | 0.188 | |||
| Chronic Kidney Disease | 2,010 | 0.51 | 0.10, 9.15 | 0.509 | |||
| Culture Requested | 6,273 | ||||||
| FALSE | — | — | — | — | |||
| TRUE | 1.60 | 1.06, 2.52 | 0.032 | 1.58 | 1.03, 2.50 | 0.043 | |
| Abbreviations: CI = Confidence Interval, OR = Odds Ratio | |||||||
Discussion
This study examined antibiotic prescribing patterns for UTI patients at KNUST Hospital, Ghana, over a 12-month period in 2023. Four principal findings emerged.
First, Access-category antibiotics accounted for only 8.0% of all antibiotic prescriptions — far below the WHO AWaRe 70% target (p < 0.001). Watch-category antibiotics, dominated by ciprofloxacin and cefuroxime, constituted 84% of prescriptions. This is consistent with the broader pattern of fluoroquinolone and third-generation cephalosporin over-reliance documented across West African health facilities, and reflects the structural embedding of these agents as empirical first-line choices. The extent of non-compliance with the WHO target suggests that antibiotic stewardship efforts at this facility must prioritise shifting prescribing toward Access agents where clinically appropriate.
Second, 81% of antibiotic encounters were purely empirical — initiated without any concurrent urine culture request. Only 19% of encounters had a culture ordered alongside the antibiotic. Importantly, culture ordering did not shift antibiotic choice toward Access agents: Watch-category prescribing remained similarly high regardless of whether a culture was ordered (85% without culture vs 79% with culture). This suggests that culture requests reflect patient complexity rather than stewardship practice, and that the results of such cultures may not be systematically fed back into prescribing decisions.
Third, treatment duration showed substantial non-concordance with the Ghana STG when sex-stratified thresholds were applied. While female patients were reasonably well-managed for the two STG first-line agents (77.6–88.7% concordance), male patients were almost universally under-treated: 95.8% of male ciprofloxacin prescriptions and 98.5% of male cefuroxime prescriptions fell below the STG-recommended 10–14 day duration. This mirrors the findings of Owusu et al. (2022), who reported that only 10% of male UTI patients in a Ghanaian primary facility received the correct duration. The 7-day prescription period appears to be the default regardless of sex, suggesting that the sex-differentiated duration recommendation from the Ghana STG is not being applied in practice.
Fourth, in the multivariable logistic regression, male sex was independently protective against Watch-category antibiotic prescribing (adjusted OR ≈ 0.53), and culture ordering was associated with higher odds of Watch-category prescribing (adjusted OR ≈ 1.58). Age, patient type, and comorbidity status were not independently associated with AWaRe category received. The culture-ordering finding is counterintuitive from a stewardship perspective and may reflect clinician tendency to order cultures in more complex presentations where broader-spectrum cover is also initiated.
Limitations include the retrospective single-centre design, which limits generalisability. Diagnoses were extracted from free-text EMR fields and may include both uncomplicated and complicated UTI without formal clinical classification. Culture results and treatment outcomes are not available in the prescribing dataset, so treatment appropriateness cannot be fully evaluated. The DDD denominator (7,507 patients) is based on registered patients rather than bed-days, which limits direct comparison with inpatient DDD benchmarks.
Conclusion
Antibiotic prescribing for UTI at KNUST Hospital, Ghana, in 2023 was dominated by Watch-category agents and fell substantially short of the WHO AWaRe 70% Access target. Empirical prescribing without culture guidance was the norm, and treatment duration for male patients was almost universally shorter than Ghana STG recommendations. These findings highlight actionable targets for antimicrobial stewardship: promoting culture ordering with systematic feedback, increasing Access-category prescribing for uncomplicated UTI where appropriate, and reinforcing the sex-differentiated duration guidance from the Ghana STG — particularly for male patients.
References
World Health Organization. AWaRe classification of antibiotics for evaluation and monitoring of use. Geneva: WHO; 2021.
Ghana Health Service. Standard Treatment Guidelines, 7th Edition. Accra: Ghana Health Service; 2017.
Owusu H, Thekkur P, Ashubwe-Jalemba J, et al. Compliance to guidelines in prescribing empirical antibiotics for individuals with uncomplicated urinary tract infection in a primary health facility of Ghana, 2019–2021. Int J Environ Res Public Health. 2022;19(19):12413. doi:10.3390/ijerph191912413
Asamoah B, Labi AK, Gupte HA, et al. High resistance to antibiotics recommended in standard treatment guidelines in Ghana: a cross-sectional study of antimicrobial resistance patterns in patients with UTIs between 2017–2021. Int J Environ Res Public Health. 2022;19(24):16556. doi:10.3390/ijerph192416556
Gupta K, Hooton TM, Naber KG, et al. International clinical practice guidelines for the treatment of acute uncomplicated cystitis and pyelonephritis in women. Clin Infect Dis. 2011;52(5):e103–e120.
World Health Organization. WHO Model Formulary. Geneva: WHO; 2023.