Background: Adherence is defined as taking medications as described or prescribed by health care professionals. Adherence is an important notion that reflects the degree to which patients conform or follow instructions and recommendations of health care providers throughout the prescribed treatment course. It involves a retrospective memory for remembering the way the medicines to be used and a prospective memory concerning the time at which the medications to be taken. The adherence process entails three main elements: initiation of therapy; implementation of the therapy as prescribed; and persistence on the given therapy for the desired period of time. Primary nonadherence refers to a situation when people do not dispense the new prescriptions written by their health care providers from the beginning of the treatment course. Although primary adherence to medications is crucial for any successful treatment strategy in both acute and chronic health conditions, less attention has been given to this issue until recently. In fact, little is known about frequency, causes, and consequences of primary nonadherence. In addition, the published literature shows variable impact of risk factors on adherence. Primary nonadherence has not been investigated previously in Hungary at the national level. Objectives: The aim of this study was to (1) estimate primary nonadherence to GMP prescribed medications among adults in Hungary using the WHO key indicator of patient care “percentage of drugs actually dispensed” as a basic concept to quantify the dispensed medications at the period between 2012 and 2015; (2) to determine effects of GMP structure and patient characteristics on adherence; (3) to describe variation of adherence across GMPs; and (4) to test whether operating GPC model for the purpose of organizing and improving effectiveness of PHC implemented in the “Public Health Focused Model Programme for Organising Primary Care Services Backed by a Virtual Care Service Centre” increases the percentage of drugs actually dispensed. Methods: National data on all GMPs were obtained from the National Health Insurance Fund and the Hungarian Central Statistical Office for the period 2012 -2015. The data were aggregated for all running GMPs around the country for all ATC group of drugs. Ratios of dispensed to written (DWRs) prescriptions written by GPs for adults aged 18 years and above were used to determine levels of primary adherence to prescribed medications at the national level. Standardized DWRs (SDWRs) were calculated using the indirect standardization method for age, sex, and eligibility for an exemption certificate. Generalized linear regression modeling was used to identify the major determinants of the SDWRs while controlling for the time. Characteristics of GMPs including patient education obtained from the GMPs, vacancy of the GMPs, type of settlement as urban or rural, county list size according to the number of adults receiving health care, and geographical location of the county were the investigated determinants in the regression model. The data analyses were completed using SPSS version 20. To evaluate impact of the SHCP, SDWRs were calculated in the first quarter of 2012 (2012Q1 representing before intervention status) and in the third quarter of 2015 (2015Q3 representing post-intervention status). Risk ratios (RR) were estimated by taking after to before ratios for SDWRs along with their corresponding 95% confidence intervals (95%CI). Results: Out of 438,614,000 written prescriptions between 2012 and 2015, 281,315,386 prescriptions were dispensed. Overall, 64.1% of the written prescriptions were dispensed. Based on the generalized linear regression coefficient (b), there was an inverse association between SDWRs and relative education of patients [b=-0.440, 95%CI: -0.468;-0.413], vacancy of the GMPs [b= -0.193, 95%CI: -0.204;-0.182], and living in urban areas [b= -0.099, 95%CI: -0.103;-0.094]. A better SDWRs was noted for GMPs running in a relatively smaller localities [bX-800= 0.052, 95%CI: 0.041; 0.063, b801-1200= 0.031, 95%CI: 0.025; 0.037, b1201-1600= 0.017, 95%CI: 0.013; 0.022] compared to those running in larger localities [b2001-X= -0.014, 95%CI: -0.019;-0.009]. In addition, geographical location of the county was an important determinant. In the intervention area where the SHCP was implemented, SDWRs indicated that overall adherence was generally higher in the intervention area than in Hungary for various ATC groups. SDWR for the entire practice was 1.042 in 2012Q1 and increased to 1.108 after the intervention in 2015Q3. When the RR was calculated for SDWRs, this change was shown to be significant [RR= 1.064; 95%CI: 1.054 - 1.073] indicating an overall improvement of 6.4% in adherence. The excess number of prescriptions dispensed was 5,033 in 2015Q3. The most significant impact observed was on both cardiovascular system drugs [RR= 1.062; 95%CI: 1.048-1.077] and alimentary tract and metabolism drugs [RR=1.072; 95%CI: 1.049-1.097] with 2,143 and 1,001 excess number of dispensed prescriptions, respectively. In addition, significant positive changes were observed for musculoskeletal drugs [RR=1.041; 95%CI: 1.010-1.074], blood and blood-forming organ drugs [RR=1.077; 95%CI: 1.044-1.111], and drugs of the nervous system [RR= 1.082; 95%CI: 1.047-1.118]. Study implications: Nonadherence contributes substantially to poor disease diagnosis. This impact was not investigated in our analysis but expected to be great since only 59.4% of cardiovascular system drugs are actually dispensed although cardiovascular diseases are the leading causes of death and the major determinants of Life expectancy in the European region. In addition, nonadherence results in considerable loss of time, work, capacities and resources of the Hungarian health care system. It may reflect poor patient-physician relationship. Also, the observed variations in adherence among various GMPs reflect various capacities of the GPs in managing and dealing with their customers. Moreover, the study provided evidence on the weak role of GPs in managing clients in urban areas and those with high levels of education. The intensive care given to patients in the SHCP was fruitful in enhancing adherence although increasing adherence was not among objectives of the programme. This probably confirms that improved patient-physician relationship is a cornerstone not only in enhancing adherence but also the overall health status of the population. Use of the WHO indicator percentage of drugs actually dispensed is a good tool for monitoring performance of GMPs and assessment of effectiveness of intervention programmes. Conclusions: About one-third of the prescriptions written by GPs working in PHC were not filled in Hungary indicating an overall alarming high rate of nonadherence. The study also demonstrated a wide variability of adherence across various GMPs. This variation can be attributed to GMPs' structural characteristics including patients’ socioeconomic status, vacancy of GMPs, list size of the county, locality type, and geographic location of the counties and more importantly magnitude of patient-physician cooperation and communication style. The SHCP provided evidence that extension of PHC services to include integrated and preventive services with proper protocol necessary capacities enhanced medication adherence. This improvement was remarkable among adult patients with cardiovascular diseases and alimentary tract and metabolic disorders. The improvement of 6.4% reported in the programme without any specific activity for increasing adherence goes in line with published studies (range 4%-11%) devoted totally to enhance adherence using multifaceted interventions. In addition, our findings proposed that DWRs can be used in routine monitoring of the operation of PHC and support substantial interventions. This finding endorses recommendations of the WHO in using the percentage of drugs actually dispensed in regular monitoring as a key indicator of patient care. Furthermore, measuring DWRs could be a useful indicator of the effectiveness of client- health care professionals’ relationships in PHC.