Association between drug shortage and factors and its duration at the market level
| Parameter or level | Shortage | Shortage duration* | ||
|---|---|---|---|---|
| OR (95% CI) | Coefficient (SE) | p value | Average marginal effect (95% CI)‡ | |
| Market structure (Ref.: B and G single) | ||||
| B and G multiple | 0.13 (0.06 to 0.25) | −0.18 (0.38) | 0.639 | −23.5 (−121 to 80.6) |
| B only | 1.15 (0.76 to 1.76) | −0.17 (0.23) | 0.465 | −22.0 (−104 to 32.0) |
| G single | 1.99 (1.28 to 3.11) | 0.51 (0.24) | 0.034 | 95.9 (11.4 to 169.0) |
| G multiple | 0.57 (0.34 to 0.97) | −0.19 (0.29) | 0.502 | −25.1 (−112 to 51.0) |
| BIO | 1.13 (0.65 to 1.97) | 0.34 (0.32) | 0.283 | 58.2 (−57.9 to 167.4) |
| Overall p value† | < 0.001 | < 0.001 | ||
| Route of administration or form (Ref.: oral solid regular release) | ||||
| Injection | 1.14 (0.82 to 1.59) | 0.05 (0.21) | 0.831 | 7.2 (−58.2 to 75.7) |
| Oral nonsolid | 1.66 (1.11 to 2.49) | 0.27 (0.21) | 0.204 | 47.1 (−22.7 to 130.1) |
| Oral solid special release | 1.05 (0.66 to 1.68) | 0.04 (0.27) | 0.896 | 5.6 (−96.0 to 150.1) |
| Other | 1.31 (0.89 to 1.91) | −0.01 (0.21) | 0.942 | −2.3 (−65.7 to 76.6) |
| Overall p value† | 0.151 | 0.758 | ||
| Proportion of DINs on formulary | ||||
| Per 10% increase | 1.03 (1.00 to 1.05) | −0.02 (0.01) | 0.071 | −3.8 (−8.3 to 0.5)§ |
| Anatomic Therapeutic Chemical Classification (Ref.: sensory organs) | ||||
| Alimentary tract and metabolism | 0.20 (0.10 to 0.38) | −0.40 (0.35) | 0.255 | −70.4 (−210 to 56.0) |
| Blood and blood-forming organs | 0.08 (0.03 to 0.22) | −0.64 (0.56) | 0.258 | −100 (−262 to 100.9) |
| Cardiovascular system | 0.51 (0.28 to 0.90) | −0.31 (0.29) | 0.281 | −57.5 (−204 to 56.1) |
| Dermatologicals | 0.36 (0.20 to 0.64) | −0.07 (0.29) | 0.803 | −14.7 (−135 to 108.8) |
| Genitourinary system and sex hormones | 0.55 (0.31 to 0.98) | −0.36 (0.30) | 0.230 | −64.2 (−196 to 46.1) |
| Systemic hormonal preparations, excluding sex hormones and insulins | 0.30 (0.15 to 0.59) | −0.41 (0.38) | 0.287 | −71.2 (−216 to 67.5) |
| Anti-infectives for systemic use | 0.43 (0.25 to 0.74) | −0.33 (0.31) | 0.287 | −59.7 (−194 to 51.9) |
| Antineoplastic and immunomodulating agents | 0.18 (0.10 to 0.34) | −0.53 (0.37) | 0.148 | −88.3 (−232 to 54.7) |
| Musculoskeletal system | 0.35 (0.17 to 0.73) | −0.01 (0.37) | 0.970 | −2.9 (−155 to 144.7) |
| Nervous system | 0.35 (0.21 to 0.59) | −0.23 (0.28) | 0.399 | −44.5 (−170 to 65.4) |
| Antiparasitic products, insecticides and repellents | 0.55 (0.10 to 2.89) | −1.03 (0.85) | 0.225 | −137 (−282 to −2.5) |
| Respiratory system | 0.58 (0.34 to 0.99) | −0.44 (0.28) | 0.111 | −75.8 (−191 to 12.6) |
| Various | 0.22 (0.11 to 0.47) | −0.18 (0.44) | 0.676 | −35.4 (−189 to 144.4) |
| Overall p value† | < 0.001 | 0.955 | ||
| Model fit statistics | ||||
| C-statistic | 0.73 | |||
| AIC | 2487.90 | 5540.4 | ||
| Log-likelihood | −1219.95 | −2745.2 | ||
Note: AIC = Akaike information criteria, B = branded manufacturer, BIO = biologic manufacturer, CI = confidence interval, DIN = Drug Identification Number, G = generic manufacturer, OR = odds ratio, Ref. = reference category (0 or 1 as appropriate), SE = standard error.
↵* Results were based on truncated negative binomial regression models among markets in shortage.
↵† Based on type III Wald Test for shortage and type III likelihood ratio test for shortage duration.
↵‡ The average marginal effect of a variable was evaluated as the average of predicted differences (from the reference group) at observed values of covariates for each observation; 95% CI were the bootstrap confidence intervals based on 2000 bootstrap samples.
↵§ For an increase from 80% to 90%.