Use of Healthcare Resources and Drug Expenditure Before and After Treatment of Chronic Hepatitis C With Direct Antiviral Agents

Mercedes Vergara; Mireia Miquel; Emili Vela; Montserrat Cleries; Caridad Pontes; Alba Prat; Montse Rué


J Viral Hepat. 2021;28(5):728-738. 

In This Article

Patients and Methods

Data Source and Study Design

A retrospective cohort analysis was performed using data from the Catalan Health Surveillance System (CHSS). Catalonia is an autonomous region in north-eastern Spain with 7,553,650 inhabitants. The public health system provides universal healthcare coverage to all residents through the Catalan Health Service (CatSalut). Detailed information on healthcare usage, including information from the minimum basic data set registered by healthcare units (eg hospitals, primary care centres, emergency departments, nursing facilities and mental health centres), is routinely collected by CatSalut.

Information on drug prescribing and some clinical parameters is collected by the invoicing system for drugs restricted for hospital use, such as DAA. Billing for services (eg outpatient visits to specialists, non-urgent medical transportation, outpatient rehabilitation, home oxygen therapy and dialysis) is also collected. Initially, the Catalan health administration deployed a series of registers to record healthcare units' activity. In 2011, the CHSS was created to integrate most of these activity registers, placing the patient (instead of the provider) at the centre of this information system, thus favouring longitudinal analysis and providing a comprehensive view of health problems across different healthcare settings and levels. Only public healthcare expenses were analysed, since information from private health centres is not connected to CHSS and thus was not available.

Patient Selection

All patients treated with DAA between 1 January 2015 and 31 December 2017 were included. All information on the use of healthcare resources from 12 months before to 24 months after stopping DAA treatment was retrieved and anonymized. The DAA treatment period was excluded from the analysis.

Data were excluded for patients who moved residence to outside Catalonia during the follow-up period, those who were followed up at private centres (all DAA treatments were from public health services), those who had not finished DAA treatment before December 2017, those who had received simultaneous treatment with DAA and interferon and those for whom severity of liver fibrosis could not be determined.

Variables analysed included number of hospital admissions (scheduled and urgent), length of hospital stay (days), admissions due to cirrhosis complications (this included any unscheduled hospital admission due to hepatic encephalopathy, ascites, spontaneous bacterial peritonitis, hepatorenal syndrome or hepatic complications such as hepatocellular carcinoma, oesophageal varices or portal hypertension), visits to the emergency department, primary care or outpatient clinics, quantity of drugs received (number of medication containers) and the exposure to different pharmacological groups, as derived from medication invoicing. The International Classification of Diseases, ninth revision, Clinical Modification (ICD-9-CM) codes were used to identify patients with cirrhosis (codes 571.2 and 571.5) (Appendix 1) as well as complications of cirrhosis.

Drug exposure was assessed through the number of invoiced units as grouped by the Anatomical Therapeutic Chemical codes specified in Appendix 2.

Healthcare resources were also analysed in different subgroups according to (1) HIV co-infection, (2) fibrosis stage at the time of starting treatment and (3) individual morbidity burden. Fibrosis stage was retrieved from the information that physicians must report in the drug invoicing system. It was reported as F0: no fibrosis, F1: minimal fibrosis, F2: moderate fibrosis, F3: severe fibrosis and F4: cirrhosis; an additional category F5: decompensated cirrhosis was created for cases in which the patient had both F4 fibrosis and (a) cirrhosis reported as "decompensated" in the invoicing system; (b) had received typical pharmacological treatment for decompensated cirrhosis (spironolactone, furosemide, carvedilol, propranolol, lactulose or lactitol) during the year prior to the start of treatment, as defined from the Anatomical Therapeutic Chemical codes specified in Appendix 2; (c) had a reported F4 fibrosis and chronic renal failure; or (d) had any urgent admission due to specific complications of cirrhosis (oesophageal varices with or without haemorrhage, hepatic encephalopathy, spontaneous bacterial peritonitis, hepatorenal syndrome, hepatocellular carcinoma, ascites or portal hypertension) as defined by ICD-9-CM codes specified in Appendix 1. Individual morbidity burden[25–27] was calculated using a population-based health risk assessment tool deployed in Catalonia, the Adjusted Morbidity Grouper (GMA), which is used to calculate an individual's morbidity burden. GMA categorizes each patient in a risk-stratification pyramid with four strata:

  • Basal risk: individuals with minimal morbidity burden; 50% of individuals in the overall population of Catalonia (but 8.1% of patients treated with DAA) fall into this stratum.

  • Low risk: individuals with low-level complexity; 30% of individuals in the overall population of Catalonia (but 30.9% of patients treated with DAA) fall into this stratum.

  • Moderate risk: higher complexity than the previous risk stratum; 15% of individuals in the overall population of Catalonia (but 43.9% of patients treated with DAA) fall into this stratum.

  • High risk: a greater morbidity burden than the previous stratum; 5% of individuals in the overall population of Catalonia (but 17.1% of patients treated with DAA).

Statistical Analysis

The descriptive characteristics of the patients (age, sex, fibrosis stage and comorbidities) correspond to those recorded at the beginning of DAA treatment. For follow-up data, all patient information was structured in monthly records, so rates of resource use per patient were calculated in periods of 30 days.

The pre-treatment period was defined as the 12 months preceding the start date of DAA treatment. The post-treatment period was defined as the 24 months following the end date of DAA treatment. The use of healthcare resources during the treatment period was excluded.

Data were summarized as percentages or means (SD). The baseline characteristics of the patients were described by sex and fibrosis stage. To estimate variation in resource use and expenditure, generalized linear mixed-effects models with a Poisson distribution were carried out, considering the patient as a random effect. The statistical analysis was carried out using the R statistical package, version 3.4.3 (lme4 library).


The study used retrospective data from administrative databases, and patients were anonymized to the researchers. The Ethics Committee at our institution (Parc Taulí) stated that under Spanish legislation informed consent was not required.