The work on this thesis began in 2003 when the global HIV epidemic was out of control. A minority of persons with HIV were benefitting fully from the recently introduced highly efficacious antiretroviral therapy (ART) combinations. Among the global challenges were lack of access to good healthcare, drug toxicity, and emergence of drug-resistant virus. It was unknown how long the drugs could maintain their efficacy in the individual even if administered as intended, and there was a fear that the increased drug pressure would increase the prevalence of drug resistance, subsequently leading to transmission of resistant virus from one individual to another, and thereby waning the treatment options available. Hence, we were far from the ideal conditions where an HIV-infected individual gets to know immediately that he/she is infected, has access to specialized medical and social support, receives a drug combination which effectively suppresses the virus and has no side effects, and is free of co-morbid conditions both before and after he/she gets infected. The nine papers on which this thesis is based each aimed to provide new knowledge to aspects of the above. Late diagnosis and late presentation to clinical care continue to be major barriers to improved HIV management. We used nation-wide hospital registries to explore the potential for an indicator disease-based HIV testing strategy. A range of conditions that were manifestations of the HIV infection itself were found to be associated with highly increased risk of HIV diagnosis during the coming year, but less so three to five years later. Other conditions were associated with an almost constant five-year long increased risk of being diagnosed with HIV because they share behavioural risk factors with HIV, making them indicators of not only current HIV but also of future HIV acquisition. Hence, indicator condition-based testing should be adapted to the local epidemic and could be a valuable addition to the existing detection practice. Once diagnosed, getting the full benefit of modern HIV care requires access to a good healthcare system. We compared temporal trends in quality and quantity of ART introduction in Den-mark and Greenland. Despite similar levels of health worker education and economic resources, ART implementation and mortality decline in Greenland lacked several years behind Denmark. The study reminded us that although economy may be a prerequisite for implementing an effective HIV care system, it is certainly not all it takes. The nationwide nature of the Danish HIV Cohort Study also allowed us to study a number of time trends at the population level. Despite what was feared, we found that the prevalence of triple-drug class virological failure (TCF) seemed to have stabilized after 2000; that the incidence rates of drug resistance acquisition were decreasing during 1999-2005; and that the prevalence of potential transmitters of drug-resistant HIV decreased during 1997-2004. We also looked at some of the consequences of virological failure and drug resistance and found that even modest levels of viraemia were associated with a high risk of future failure and death, and that in persons who have experienced TCF, the number and pattern of resistance mutations were independent predictors of death. Hence, despite the overall positive trends in virological failure and drug-resistance development at the population level, our findings underline the crucial importance of always having an effective treatment option available for the individual patient with drug-resistant virus. As mortality was declining for persons with access to ART and good HIV care, it became important to know how long persons with HIV could expect to live compared to the general population. We projected long-term survival and found that a 25-year old person with HIV and without hepatitis C virus (HCV) co-infection had a 50 per cent chance of surviving another 39 years, only 12.2 years less than a person in a matched general population cohort would survive. With improved survival and declining HIV-related co-morbidity, non-HIV related co-morbidity became a more visible contributor to the health status of persons with HIV. We assessed the impact of non-HIV related comorbidity acquired before the person became infected with HIV. We found that 32% of the observed mortality in our cohort was due to HCV and co-morbidities measured by the Charlson Comorbidity Index, 13% corresponded to the background mortality in the population, and that only 55% of the mortality could be attributed to HIV. Our findings confirmed that persons acquiring HIV differ at large from the general population, and that we should not expect overall mortality rates in populations with HIV to reach the levels in the general population. This thesis attempted to map some of the many challenges on the road towards increased survival of individuals and populations with HIV up to a level, which today in many settings is close to that of the general population. The studies in this thesis have each paid their modest contribution to show how crucially important it is to be diagnosed in time, to have access to a well-functioning health system, and to keep free of co-morbidity both before and after acquiring HIV. After many years of struggle and despair, and thanks to enormous advances in prevention and treatment, we are now looking towards a promising future.