Real-world healthcare resource utilisation for migraine: Analysis of an administrative database in Colombia

Abstract

Introduction

Migraine is a chronic condition characterised by recurrent episodes of severe headache, significantly affecting the functionality of the individuals. The global prevalence and the impact on quality of life highlight the need for research on its epidemiology and management, especially in specific contexts such as Colombia, where information on the subject is scarce.

Methods

A retrospective descriptive study was conducted with data from the Information System for Social Protection in Colombia (SISPRO) between 2027 and 2021, evaluating the incidence, prevalence, treatment pattern and cost associated with migraine management.

Results

293 028 incident cases and 1 711 794 prevalent cases of migraine were identified with a higher incidence and prevalence in women. The use of prophylactic treatments and visits to the neurology department showed an increasing trend. Although pharmaceutical costs represented the greatest economic burden, they decreased over the course of the study. The annual healthcare cost per patient, expressed in purchasing power parity (PPP) decreased from $110 PPP in 2017 to a minimum of $39 PPP in 2019. The annual pharmacy cost per patient also dropped from $115 PPP in 2017 to $64 PPP in 2019.

Conclusions

This study highlights migraine as a common condition in Colombia, with a prevalence and management that reflect a need for improved educational and treatment strategies. The research also highlights the importance of national data in understanding and addressing this condition, suggesting the need for public health policies focused on the prevention and optimised management of migraine.

Resumen

Introducción

La migraña es una condición crónica caracterizada por episodios recurrentes de cefalea intensa, afectando significativamente la funcionalidad del individuo. La prevalencia global y el impacto en la calidad de vida resaltan la necesidad de investigaciones sobre su epidemiología y manejo, especialmente en contextos específicos como Colombia, donde la información al respecto es escasa.

Métodos

Se realizó un estudio descriptivo retrospectivo con datos del Sistema de Información para la Protección Social en Colombia (SISPRO) entre 2017 y 2021, evaluando la incidencia, prevalencia, patrones de tratamiento y costos asociados al manejo de la migraña.

Resultados

Se identificaron 293.028 casos incidentes y 1.711.794 casos prevalentes de migraña, con una mayor incidencia y prevalencia en mujeres. La utilización de tratamientos profilácticos y las consultas a neurología mostraron una tendencia creciente. Aunque los costos farmacéuticos representaron la mayor carga económica, disminuyeron a lo largo del estudio. El costo sanitario anual por paciente, expresado en Paridad del Poder Adquisitivo (PPA), disminuyo de PAA$110 en 2017 a un mínimo de PAA$39 en 2019. El costo anual de farmacia por paciente también se redujo de PAA$115 en 2017 a PAA$64 en 2019.

Conclusiones

Este estudio destaca la migraña como una condición frecuente en Colombia, con una prevalencia y manejo que reflejan una necesidad de estrategias educativas y de tratamiento mejoradas. La investigación también subraya la importancia de los datos nacionales en la comprensión y abordaje de esta condición, sugiriendo la necesidad de políticas de salud pública enfocadas en la prevención y manejo optimizado de la migraña.

Introduction

Migraine is a chronic disorder characterised by recurrent episodes of intense headache, frequently pulsatile, accompanied by a wide range of symptoms, which may significantly limit daily functioning.¹ Its worldwide prevalence has recently been estimated at 14.4%, with higher rates among women (18.9%) than men (9.8%).² In Colombia, migraine prevalence is 13.8% among women and 4.8% among men.³

Migraine is classified according to different criteria: by phenotype, as migraine with or without aura; and by frequency, as episodic (EM) or chronic migraine (CM). It should be noted that CM involves greater disease burden and is associated with greater analgesic use and higher levels of disability.¹

Pharmacological treatments for migraine are classified as acute or prophylactic. Acute treatment includes such agents as acetaminophen, acetylsalicylic acid, non-steroidal anti-inflammatory drugs (NSAID), triptans, ergot derivatives, gepants, and ditans.⁴ Prophylactic treatment includes beta-blockers, neuromodulators, tricyclic antidepressants, calcium channel blockers, botulinum toxin, gepants, and calcitonin gene–related peptide monoclonal antibodies.⁵

Migraine has a significant impact on quality of life, with up to 25% of patients missing one day of work per month and 54% reporting significant functional impairment during severe attacks.⁶ It is also estimated to cause a 43.4% health loss compared to the healthy population, according to the Global Burden of Disease Study 2016. Likewise, migraine is associated with significant healthcare resource utilisation (HCRU).⁷ According to recent studies, patients with CM require a mean of 5.38 primary care consultations and 1.9 specialist consultations per semester, with 7% requiring hospitalisation.⁸ From an economic viewpoint, the largest proportion of costs is attributed to lost productivity, estimated at $5 billion in the United States. Costs associated with CM are 2-4 times greater than those for EM.⁹

Despite its relevance, no study has analysed migraine epidemiology and HCRU in Colombia. The present study addresses the need for real-world data from our setting. To this end, we conducted a comprehensive analysis of Colombia’s Integrated Information System for Social Protection database (SISPRO, for its Spanish initials), which collects reports from all healthcare providers in the Colombian public healthcare system, covering 99% of the population of Colombia.

Material and methods

Design

We conducted a descriptive study of retrospective clinical data from SISPRO between 2017 and 2021. As the database provides aggregated data, the study was designed as a cross-sectional trend analysis.

Data sources

Epidemiological and HCRU data were extracted from SISPRO according to the International Classification of Diseases (ICD-10) code. In the context of this study, HCRU data encompassed medical procedures, medications, hospital admissions, and outpatient visits recorded for patients with a diagnosis of migraine.

SISPRO is an official government database that collects healthcare data to support clinical decision-making. This structured database collects reports from all healthcare providers within the Colombian public healthcare system, which covers approximately 99% of the population. The comprehensiveness of the database enables temporal analyses including nearly the entire Colombian population with a given disease.

Several independent modules within SISPRO differentiate between data from outpatients, inpatients, and pharmacological interventions. One of these modules is the Individual Register of Healthcare Services Provision (RIPS, for its Spanish initials). Data in this module are reported by healthcare providers as a billing record for the healthcare services provided. It includes information on consultations, procedures, emergency care, and hospitalisations. All services are coded using the national classification system for medical procedures, known as the Single Code for Healthcare Procedures (CUPS, for its Spanish initials). The module also records the diagnostic code associated with each service according to the applicable version of the ICD.

For cost estimation, data were gathered from the Information System for Drug Prices (SISMED, for its Spanish initials), as recommended by the Agency for Health Technology Assessment (AETS). Data from the prescriptions module (MIPRES) are linked to the costs obtained from SISMED using the Single Code for Medicinal Products, a coding system that assigns a unique code to each drug presentation present in both databases. MIPRES is an electronic platform used by healthcare professionals in Colombia to prescribe healthcare technologies and services that are not included in the Colombian healthcare service’s drug list.

Each diagnosis recorded in SISPRO is classified into one of 3 categories: 1) new, confirmed; 2) repeated, confirmed; or 3) working diagnosis. With a view to obtaining accurate information about incidence and treatment patterns, we only included diagnoses labelled as “new, confirmed.” This criterion, combined with unique patient identification numbers, ensures data integrity, as it prevents duplications in the assessment of epidemiological parameters over time.

Patients

We included all patients diagnosed with migraine (whether episodic or chronic) according to the ICD-10, based on data available in the SISPRO databases during the study period.

The inclusion criteria were age ≥ 18 years, known place of residence, and any of the following ICD-10 diagnoses: migraine without aura (G43.0), migraine with aura (G43.1), status migrainosus (G43.2), complicated migraine (G43.3), other migraine (G43.8), and unspecified migraine (G43.9).

Study variables

With a view to establishing cross-sectional estimations of epidemiological measures and HCRU, the following variables were retrospectively gathered from the SISPRO databases: sex, age, type and day of consultation, hospitalisation and its duration, procedures, medications prescribed and covered by the national healthcare plan, units per medication, and healthcare costs.

Statistical analysis

A descriptive analysis was performed of the data gathered, using measures of central tendency and dispersion. The results are presented in tables and figures. Continuous variables are expressed as means or medians, with dispersion expressed as variance and range. Categorical variables are expressed as absolute values or relative frequencies. We also calculated 95% confidence intervals (CI).

For the purposes of this study, annual prevalence was defined as the number of individuals living with a given condition within a given year, whereas annual incidence was defined as the number of new diagnoses of that condition within a given year. It should be noted that prevalence and incidence estimates were based on data from consultations that took place during the study period. HCRU considered medical consultations, dispensed medications, diagnostic tests, and frequency of hospitalisation. Pharmacological costs were analysed by medication type (prophylactic or acute) and by subtype within each category.

Given that the information gathered corresponds to aggregated data for the total population by ICD-10 code, and patient identifiers were not available, no patient follow-up or analysis of individual treatment patterns was performed.

Results

During the study period (2017-2021), a total of 293 028 incident cases of migraine were identified in Colombia. It should be noted that we observed an incidence of 230.9 cases per 100 000 population in 2019, which was significantly higher than in previous and subsequent years (data not shown; Supplementary Table 1). Regarding prevalence, a total of 1 711 794 cases of migraine were identified during the study period. As with incidence, prevalence was higher in 2019, at 1694.1 cases per 100 000 population, with an age-standardised prevalence rate of 1742.3 cases per 100 000 population (data not shown, Supplementary Table 2). Furthermore, the use of prophylactic medications and the number of neurology consultations showed an upward trend over the study period. The number of primary care consultations was higher at the beginning of the study period.

Incidence

Interestingly, the crude and the age-standardised incidence rates of migraine were lower in 2020 and 2021 than in previous years. By age group, the highest incidence was observed among patients aged 25-29 years, followed by the 30-34 year age group (Table 1). By sex, migraine incidence was up to 4 times higher in women than in men, with rates of 221.4-360.6 cases per 100 000 population in women and 51.7-85.2 cases per 100 000 population in men (Supplementary Table 1). This trend persisted across all age groups up to 65 years, after which the difference began to diminish (Table 1 and Supplementary Table 1).

Table 1.

Incidence of migraine per 100 000 population: overall and stratified by sex, age group, and year.

Year	2017		2018		2019		2020		2021
Sex	Women	Men	Women	Men	Women	Men	Women	Men	Women	Men
Age group
18-19 years	252.0	55.5	278.1	68.3	359.5	89.2	200.7	51.8	224.0	47.3
20-24 years	336.7	75	357.0	85.1	469.3	111.2	281.7	65.6	304.8	73.5
25-29 years	370.4	78.6	389.2	92.9	514.3	123.8	318.9	75.4	344.2	76.2
30-34 years	358.5	67.7	385.5	80.1	502.3	111.3	313.6	65.5	333.4	67.1
35-39 years	332.9	66.3	361.0	68.6	467.3	95.7	289.7	58.6	308.7	62.5
40-44 years	304.5	52.4	325.4	62.2	437.4	83.6	269.6	52.1	293.2	54.0
45-49 years	283.0	49.8	296.7	52.8	402.4	73.9	245.9	47.4	251.5	46.4
50-54 years	229.4	48.1	252.9	47.9	318.3	67.4	198.9	45.0	208.0	42.5
55-59 years	169.5	42.5	179.1	44.3	242.2	61.9	154.7	39.4	156.4	35.4
60-64 years	117.3	34.8	128.4	38.0	169.3	55.3	111.6	34.1	109.3	32.6
65-69 years	94.2	37.3	101.6	41.3	141.0	52.3	86.9	30.8	82.0	32.4
70-74 years	71.9	35.1	73.5	38.1	97.6	46.9	64.8	27.2	64.5	26.1
75-79 years	52.5	33.1	63.8	39.9	76.2	47.7	54.6	30.1	54.9	27.9
≥ 80 years	39.2	23.1	41.4	29.9	48.7	37.8	35.0	19.2	37.2	22.7
Total	260.4	56.6	385.5	68.3	359.5	89.2	200.7	51.8	224.0	47.3

Regarding incidence by health insurance plan, no specific trends were observed over the study period, although the subsidised scheme presented higher rates than the contributory scheme in 2017, 2020, and 2021, and the opposite was observed in the remaining years (Supplementary Table 1). By region, the Amazon region presented the highest rate, whereas the lowest rate was observed in the Insular region. Interestingly, the incidence rate in the Orinoco region peaked between 2018 and 2019, with higher rates than in the remaining regions (data not shown).

Prevalence

The highest prevalence rate was observed in 2019, while the lowest rate was observed in 2020, with a crude rate of 1066.0 cases and an age-standardised rate of 1035.5 cases per 100 000 population (Fig. 1 and Supplementary Table 2). Similarly to incidence, migraine prevalence was higher among women aged up to 65 years, which reflects greater susceptibility to this condition in women than in men (Table 2).

Figure 1.

Annual prevalence of migraine (cases per 100 000 population), by health insurance plan and sex.

Table 2.

Prevalence of migraine per 100 000 population: overall and stratified by sex, age group, and year.

Year	2017		2018		2019		2020		2021
Sex	Women	Men	Women	Men	Women	Men	Women	Men	Women	Men
Age group
18-19 years	1741.2	403.9	1892.3	478.6	2448.3	620.7	1470.6	365.2	1668.6	395.7
20-24 years	2632.8	688.7	2843.2	777.3	3509.1	970.5	2112.0	546.3	2505.0	657.8
25-29 years	3018.9	772.5	3249.9	865.8	3999.2	1100.0	2388.7	617.5	2852.8	725.7
30-34 years	2794.1	650.7	3038.4	745.2	3797.7	943.0	2304.1	528.4	2757.5	627.9
35-39 years	2449.7	526.0	2711.4	589.2	3414.8	761.6	2086.3	438.9	2510.8	522.0
40-44 years	2132.4	414.6	2377.1	476.1	3077.9	620.1	1944.3	374.6	2300.7	435.3
45-49 years	1878.2	352.9	2055.6	389.4	2680.6	514.9	1721.1	321.1	1996.4	366.6
50-54 years	1509.7	313.5	1668.8	345.8	2183.1	464.5	1426.7	287.8	1630.3	320.6
55-59 years	1100.1	266.2	1206.7	298.7	1626.7	402.3	1080.4	268.4	1217.3	289.8
60-64 years	758.5	231.2	852.9	253.6	1164.3	350.6	783.5	240.5	896.5	244.1
65-69 years	599.0	214.8	647.8	226.2	894.0	313.6	621.8	215.9	676.7	214.6
70-74 years	469.8	213.1	500.2	220.9	670.9	289.7	471.3	209.0	490.4	207.0
75-79 years	374.1	193.3	398.6	224.9	547.3	294.5	373.6	210.5	392.5	187.4
≥ 80 years	248.6	168.7	294.5	196.6	402.9	255.4	263.6	172.8	256.3	157.1
Total	1911.4	463.1	2076.7	518.9	2628.0	667.6	1624.2	394.5	1895.7	450.7

Unlike incidence, prevalence varied significantly according to the health insurance plan, with the contributory scheme showing consistently higher rates than the subsidised scheme (Fig. 1). By region, prevalence rates increased consistently and systematically in all regions in 2019, and decreased thereafter, except in the Insular region. Furthermore, this region consistently presented the lowest prevalence rates over the study period, whereas the region of Bogotá presented the highest prevalence during the same period (data not shown).

Patterns of medication use

In 2017, 65.1% of patients received acute treatments, mainly ergotamine. By 2021, the use of acute medications had increased significantly to 81.5%, with non-steroidal anti-inflammatory drugs being the most frequently prescribed drugs (36.8% of prescriptions). On the other hand, opioids and triptans were the least frequently prescribed; the highest prescription rates were observed in 2017 for opioids (5.5%) and in 2021 for triptans (4.3%).

The use of prophylactic medications increased progressively over the study period, from 24.1% in 2017 to 40.0% in 2021 (Table 3). In 2017, propranolol was the most frequently used drug for migraine prophylaxis (12%). The use of botulinum toxin also increased significantly, with the prescription rate in 2021 being twice as high as that observed in 2017. In 2021, a total of 49 598 units were dispensed, corresponding to 2.9% of all patients under treatment (Table 3). It should be noted that the dose of botulinum toxin was calculated based on aggregated SISPRO data, which do not include patient-level information. In Colombia, botulinum toxin is available in vials of 50, 100, 200, and 500 units; SISPRO only provides data on the number of vials used per size. Furthermore, erenumab and galcanezumab were introduced for migraine treatment in 2021, although they were used by less than 1% of the cohort.

Table 3.

Number of units of acute and prophylactic treatment dispensed per patient diagnosed with migraine, by study year.

	2017				2018				2019
	Dispensed units	Patients receiving each medication	Units dispensed per patient	% of patients receiving medication	Dispensed units	Patients receiving each medication	Units dispensed per patient	% of patients receiving medication	Dispensed units	Patients receiving each medication	Units dispensed per patient	% of patients receiving medication
				N = 398 368				N = 445 435				N = 580 694
Acute treatment	4 488 290	259 356	17.3	65.1	3 923 778	241 896	16.2	54.3	5 073 295	324 996	15.6	55.97
Ergotamine	3 389 999	109 690	30.9	27.5	2 740 681	85 112	32.2	19.1	3 796 183	141 296	26.9	24.33
NSAIDs^a	718 854	82 918	8.7	20.8	767 971	82 155	9.3	18.4	828 704	99 424	8.3	17.12
Steroids	66 532	41 189	1.6	10.3	77 161	46 496	1.7	10.4	85 230	55 084	1.5	9.49
Opioids^a	88 961	21 929	4.1	5.5	87 362	20 575	4.2	4.6	81 530	17 095	4.8	2.94
Triptans	223 944	3630	61.7	0.9	250 603	7558	33.2	1.7	281 648	12 097	23.3	2.08
Prophylactic treatment	5 529 836	95 927	57.6	24.1	4 533 711	106 552	42.5	23.9	8 787 327	179 172	49.0	30.85
Propranolol	2 028 700	47 756	42.5	12	1 767 879	41 249	42.9	9.3	4 738 745	89 849	52.7	15.47
Topiramate	1 602 365	6416	249.7	1.6	477 631	9882	48.3	2.2	1 074 593	18 486	58.1	3.18
Botulinum toxin	31 999	4414	7.2	1.1	32 827	6148	5.3	1.4	45 790	8511	5.4	1.47
100 U	10 476	808	24.5	0.2	6117	1134	5.4	0.3	6327	1450	4.4	0.2
200 U	21 092	3494	13.3	0.9	25 245	4815	5.2	1.1	38 460	6854	5.6	1.2
300 U	4	2	2.0	0.0	11	4	2.8	0.0	13	6	2.2	0.0
500 U	66	42	3.0	0.0	78	28	2.8	0.0	63	31	2.0	0.0
50 U	361	68	17.3	0.0	1376	167	8.2	0.0	927	170	5.5	0.0
Erenumab	–	–	NA	0.0	–	–	NA	0.0	–	–	NA	0.0
Flunarizine	104 794	879	119.2	0.2	369 341	9771	37.8	2.2	823 511	18 301	45.0	3.15
Valproic acid	1 242 144	17 129	72.5	4.3	1 354 173	16 321	83.0	3.7	1 424 300	18 040	79.0	3.11
Galcanezumab	–	–	NA	0.0	–	–	NA	0.0	–	–	NA	0.0
Amitriptyline	487 835	14 919	32.7	3.7	499 033	17 033	29.3	3.8	634 598	17 474	36.3	3.01

	2020				2021
	Dispensed units	Patients receiving each medication	Units dispensed per patient	% of patients receiving medication	Dispensed units	Patients receiving each medication	Units dispensed per patient	% of patients receiving medication
				N = 363 462				N = 427 091
Acute treatment	2 668 347	196 609	13.6	54.1	4 498 518	348 082	12.9	81.5
Ergotamine	1 255 833	55 535	22.6	15.3	2 591 628	120 005	21.6	28.1
NSAIDs^a	1 129 278	101 320	11.1	27.9	1 517 745	157 147	9.7	36.8
Steroids	46 846	21 323	2.2	5.9	85 903	44 724	1.9	10.5
Opioids^a	64 013	6968	9.2	1.9	66 943	7633	8.8	1.8
Triptans	172 377	11 463	15.0	3.2	236 299	18 573	12.7	4.3
Prophylactic treatment	3 738 485	100 583	37.2	27.7	8 499 564	170 802	49.8	40.0
Propranolol	1 272 355	34 101	37.3	9.4	4 962 261	72 435	68.5	17.0
Topiramate	718 016	13 557	53.0	3.7	1 183 325	22 702	52.1	5.3
Botulinum toxin	38 902	9479	4.1	2.6	49 598	12 536	4.0	2.9
100 U	6977	1408	5.0	0.4
200 U	31 712	7995	4.0	2.2	41 254	10 563	3.9	2.5
300 U	11	8	1.4	0.0	20	12	1.7	0.0
500 U	46	28	1.6	0.0	13	8	1.6	0.0
50 U	156	40	3.9	0.0	361	82	4.4	0.0
Erenumab	–	–	NA	0.0	1598	322	5.0	0.1
Flunarizine	618 526	16 362	37.8	4.5	844 355	24 174	34.9	5.7
Valproic acid	801 148	9400	85.2	2.6	783 672	13 299	58.9	3.1
Galcanezumab	–	–	NA	0.0	1246	241	5.2	0.1
Amitriptyline	250 636	8205	30.5	2.3	623 911	12 557	49.7	2.9

Unit: unit/minimum quantity of dispensed medication.

NA: not applicable; NSAID: non-steroidal anti-inflammatory drug.

a

Combination medications containing non-steroidal anti-inflammatory drugs and opioids are counted in both groups; however, only a small proportion of patients received these medications.

Medical consultations

During the study period, primary care physicians were the main healthcare providers attending patients with migraine. However, an upward trend was observed in the number of consultations with neurology departments, increasing from 58.2 consultations per 1000 patients in 2017 to 153 in 2021. Other specialties frequently consulted by patients with migraine were internal medicine, family medicine, psychology, ophthalmology, and physical therapy (data not shown).

Complementary tests

CT was the most frequently used imaging technique, before MRI. Its use peaked in 2018 (15.6 studies per 1000 patients). The highest rate of MRI use was observed in 2020, at 9.43 studies per 1000 patients (data not shown).

Costs

The analysis of medication-related costs revealed interesting trends. Although prophylactic medications were prescribed less frequently, their cost was significantly higher (Table 4). However, over the study period, a decrease was observed in the costs of prophylactic medications despite increases in prescription rates; this may reflect lower drug prices or more efficient use. In contrast, the costs of acute medications remained constant despite increases in prescription rates; this suggests that prices did not vary significantly over the study period (Table 4).

Table 4.

Mean annual healthcare cost and pharmacological cost, expressed in purchasing power parity dollars, per patient diagnosed with migraine in Colombia.

Year	2017	2018	2019	2020	2021
Mean total annual healthcare cost
Medical consultations ($ PPP)	40 030 715	24 924 051	16 284 010	12 849 838	13 184 848
Hospitalisation ($ PPP)	383 066	720 729	1 052 433	912 229	1 006 238
Diagnostic tests ($ PPP)	3 350 808	6 163 630	5 339 588	3 476 393	4 046 300
• CT ($ PPP)	384 810	922 625	1 034 058	749 185	867 215
• MRI ($ PPP)	429 671	813 323	960 598	950 290	971 731
Total annual healthcare cost ($ PPP)	43 764 589	31 808 410	22 676 031	17 238 459	18 237 386
Annual healthcare cost per patient ($ PPP)	110	71	39	47	43
Pharmacy costs
Total annual pharmacy cost ($ PPP)	45 967 362	33 205 084	37 145 359	29 172 880	31 342 997
• Acute treatments ($ PPP)	8 112 578	6 649 659	5 269 426	4 875 875	8 004 484
• Prophylactic treatments ($ PPP)	37 854 784	26 555 425	31 875 933	24 297 005	23 338 514
Total annual pharmacy cost per patient ($ PPP)	115	75	64	80	73

PPP for the years 2017−2021: 1 Colombian peso/$ PPP in 2017: 1 328 017; 2018: 1 322 161; 2019: 1 330 757; 2020: 1 297 100; 2021: 1 353 383.

CT: computed tomography; MRI: magnetic resonance imaging; PPP: purchasing power parity.

Pharmacological treatments were found to represent the largest proportion of the total costs of migraine management, followed by medical consultations and hospitalisation. Interestingly, the year 2019 presented the lowest total cost per patient, although this was not the case for each individual component (Table 4).

In our cost analysis, we used the purchasing power parity (PPP) indicator to allow comparison between countries, adjusting for differences in price levels. The annual healthcare cost per patient was found to decrease from $110 PPP at the beginning of the study period to $39 PPP in 2019. Similarly, the annual pharmacy cost per patient decreased from $115 PPP in 2017 to $64 PPP in 2019, increasing slightly to $73 PPP in the final year of the study period.

Discussion

This study represents a significant milestone in the understanding of migraine management in Colombia, as it is one of the first to analyse HCRU and treatment costs using SISPRO data. During the study period, migraine prevalence ranged from 1066.0 cases per 100 000 population in 2020 to 1742.3 cases per 100 000 population in 2019. These estimates are consistent with those reported in the ESENCIA study, which analysed the period 2015-2017, reporting a prevalence of 2170 cases per 100 000 population.¹⁰ However, they are lower than the rates reported in the Global Burden of Disease Study 2016, which estimated an age-standardised prevalence of 13 000-14 000 cases per 100 000 population, as well as the rates reported in other previous studies.²

These results suggest that migraine is underdiagnosed, possibly due to difficulty accessing healthcare, as reported previously.¹¹ Furthermore, the necessary diagnostic procedures may not be performed, and the specialists attending these patients may not be applying all diagnostic criteria. Therefore, it is essential to continue expanding health services for patients with migraine, developing and implementing efficient, dynamic care models able to overcome barriers to access. This, in turn, would optimise the standardisation of migraine management, improve treatment adherence, reduce hospitalisation rates, and improve patient quality of life.

Furthermore, the discrepancies observed in the prevalence rates reported² may be attributed to several factors. First, methodology varies considerably between studies; for example, in the Global Burden of Disease Study, rates are adjusted based on published estimates. Second, we included patients who sought medical attention, thereby excluding those who self-medicate due to low disease burden (eg, low-frequency episodic migraine) and are consequently not recorded in the reviewed databases. Third, patients diagnosed with migraine may have been coded with nonspecific ICD-10 diagnoses (eg, headache), and we may as a result have excluded them from our analysis to maintain specificity in the assessment of other outcomes, such as treatment utilisation and costs.

The age distribution of our cohort was similar to those of previous studies, confirming that migraine is most frequent in young adults.² Sex-based analysis revealed that migraine prevalence was up to 4 times higher in women than in men; this difference diminished with age. Our results are consistent with the evidence reported to date.¹² The decrease in migraine incidence in 2020-2021 observed in our study may be attributed to a decrease in the use of outpatient services during the COVID-19 pandemic.^13,14

Our results show significant regional variability across the country. The Amazon region presented a high incidence rate compared to other regions, whereas rates in the Orinoco region presented fluctuations and the Insular region showed the lowest rates. These results are consistent with those of previous studies, such as the ESENCIA study, which reported regional differences in migraine burden.¹⁰ Such disparities may be attributed to a number of factors, including genetic predisposition,⁷ environmental influences, and inequalities in access to healthcare.^15,16

Furthermore, Bogotá consistently showed the highest prevalence rate, whereas the Insular region presented the lowest; this may reflect underreporting or underdiagnosis in certain areas.¹⁷ These findings underscore the importance of considering the local context and disparities in healthcare in the management of patients with migraine.

Regarding treatment patterns, our cohort showed lower rates of prophylactic drug prescription than previous studies in other countries, such as the United States, where the prescription rate of prophylactic medications in 2018 was 8 percentage points higher than in our study. However, prescription rates for antiseizure medications and beta-blockers are comparable, and strikingly high.¹⁸

Furthermore, the low level of triptan use observed in our study is consistent with the data reported in a recent retrospective study conducted by a health maintenance organisation in Colombia. That study covers years beyond the end of our study period (2018-2022), and shows low triptan prescription rates, starting at 4% and increasing progressively to a maximum of 16%.¹⁹

In contrast, the use of botulinum toxin in the United States is significantly higher than in our cohort (13% vs 1%-3%).¹⁸

The introduction of erenumab and galcanezumab in 2021 expanded the available options for migraine prophylaxis. Use of these drugs is associated with greater treatment adherence and persistence, and lower rates of discontinuation due to adverse events, as shown in studies conducted in the United States.²⁰ Regarding acute treatment, a high proportion of patients in our cohort were prescribed acute medications, with prescription rates 15% higher than in the United States.¹⁸ This highlights the need to improve the prophylactic management of these patients to prevent such complications as medication overuse headache.

Primary care physicians play an essential role in migraine diagnosis and treatment, as is also reflected by our results.^21,22 However, we did observe an upward trend in the number of neurology consultations, which reflects growing recognition of the need for specialised knowledge to provide comprehensive care for patients with migraine. Efficient interdisciplinary cooperation improves care quality and ensures optimal outcomes.²¹

According to our results, CT is used considerably more frequently than MRI, underscoring the need to review diagnostic practices in patients with migraine and consider the appropriateness of imaging studies.

In our study, pharmacy costs represent the largest proportion of total migraine management costs. This is consistent with reports from the United States, where medication-related costs represent 80% and 70% of total healthcare cost for CM and EM, respectively.²³ Similar results have been reported in Latin America, where the direct costs of CM are comparable to those of such conditions as diabetes or hypertension, confirming that pharmacological costs represent the largest contributor to migraine-related expenditure.²⁴ Prophylactic medications also represented a significant proportion of total pharmacological costs, underscoring the importance of long-term preventive strategies. Our data revealed a downward trend in pharmacological spending over the study period, which suggests a decrease in drug prices and an improvement in affordability. These findings underscore the importance of considering all associated costs when assessing the economic burden of migraine, rather than focusing solely on prescription frequency.

One limitation of this study is the data source, which consists of national databases including data reported by healthcare providers. These may contain coding errors, data entry errors, and incomplete or inconsistent information, which may result in measurement errors within SISPRO. Regarding costs, we observed significant differences in outpatient consultation expenses between 2017 and other years of the study period; this trend was observed for all ICD-10 diagnoses. This may reflect changes in healthcare reporting policies or limitations of the data source. Furthermore, only aggregated data were available; this prevents detailed, patient-level analysis, such as patient follow-up. Given that, unlike CM, EM does not have a specific ICD-10 code, data must be analysed for prolonged migraine as a whole. Finally, procedures and consultations are not linked to a specific diagnosis; therefore, the resources used to treat comorbidities may affect HCRU estimates for any specific ICD-10 code.

Conclusions

Our results underscore the importance of managing migraine as a condition that imposes a high economic burden on the Colombian healthcare system.

Differences in prevalence and incidence rates between regions and sexes also emphasise the need for personalised, patient-focused strategies, as well as the need for public health policies that account for local factors, including healthcare accessibility and equity of healthcare resource distribution. Suboptimal migraine management, reflected by the low prescription rates for prophylactic medications, points to an urgent need to improve migraine education and training among primary care physicians, who are the main care providers for these patients. Greater specialisation and referral to the neurology department may significantly improve treatment quality and reduce dependency on acute treatment. The implementation of specific training and prevention programmes for each region may help to reduce inequalities and to improve the care provided to these patients. Furthermore, the economic burden of migraine, which results mainly from high medication-related costs, underscores the need to optimise resource use. This includes promoting more effective and accessible prophylactic treatments, which may in turn reduce long-term costs and improve patient quality of life.

CRediT authorship contribution statement

BMT and ZMM: study concept and methodology. IQVIA: study design, implementation, and statistical analysis. ZMM and BMT: data interpretation. SF, RML, ZMM, and BMT: manuscript drafting, revision, and editing. All authors had unrestricted access to the study data and granted their consent to send the manuscript for publication.

Funding

This study received financial support from Abbvie. Abvvie participated in data interpretation and in manuscript revision prior to approval. IQVIA and Biopress acted as consultants for Abbvie and received financial support from Abbvie for this study. None of the authors received honoraria.

Acknowledgements

The authors thank Biopress (Madrid, Spain) and German Cantillo-Mackenzie and Tatiana Pacheco (Medical Department, Abbvie Cluster North-LATAM; Bogotá, Colombia) for their technical support and medical writing services.