Health care services under pressure - Consequences for patient flows, efficiency and patient safety

The purpose of this project was to investigate effects of capacity pressure in health services. Such capacity pressure may overwhelm local treatment capacity and impair patient treatment, increase health care service level shifts, and health care workers may experience temporarily increased workloads. If hospitals respond to pressure by discharging patients earlier, or general practitioners by referring more patients to specialized services, strain in one part of the health care system may cause strain elsewhere. Studies of the consequences of capacity pressure in health services are rarely amenable to randomized trials, both for practical and ethical reasons. It?s also challenging to conduct studies with observational data in a health service described by incomplete data. We aimed to use principles from randomized trials to analyze observational data to address such methodological challenges. In this way we may subject organizational changes to requirements of empirical evidence, just as with the introduction of new drugs. Generally, our results suggest that the Norwegian health services adequately handle busy situations. We find no general effects of capacity pressure on patient outcomes such as mortality, readmissions, and health service use, but there are clear signs that the service is adapting when under strain. At the same time, we have shown that some patient groups may be at risk. In published work from this project, we have mostly used register data, but also qualitative interviews and health surveys. - In a qualitative study of general practitioners' work situation, concerns were expressed regarding capacity pressure and patient safety. However, there was no uniform description of the mechanisms that lead to increased workload. - A general physician's decision to refer a patient has a large impact on hospital costs via the costs of subsequent hospital treatment. Compared with the quarter that referred the least, 30-days acute hospital costs was around NOK 4.2 million higher per 1,000 out-of-hours visits by elderly patients in the quarter who referred the most. At the same time, these physicians seemed to refer more patients who subsequently received a serious diagnosis in hospitals. - We studied general practitioner consultations during a ten-year period after participating in the Nord-Trøndelag Health Survey (2006-08). There was a general increase in consultations, most pronounced among the healthiest. - We examined the use of general practitioners the year before hospitalization and found a substantial increase in contacts in the time before admission. However, many patients had no contacts the month before admission, which may indicate a potential for prevention. - We have used detailed information about the time of admission and the time of reported arrival to predict hourly arrivals at a hospital?s emergency department. This can be used to improve predictability in busy situations. - We studied length of stay in the emergency department and found no association between delays due to busyness and risk of death. However, the extra time spent in the department increased the chance that the patient was able to return home without hospitalization. - We studied how the admission time of hip fracture is associated with 30-day mortality. There were small differences in mortality associated with the day of admission. However, time of the day of a hospitalization was associated with mortality. - There have been concerns about the consequences of reduced length of stay. We used daily fluctuations in discharges to assess the hip fracture patient's risk of discharges from organizational causes. The result indicates a substantial increased mortality from organizational discharges among hip fracture patients. - We used information about concurrent surgical patients to identify situations with possible strain on resources. Hip fracture patient admitted during periods with high volumes of recent surgical patients had longer time to surgery and higher mortality. We found weak or no such effect in surgical patients with cardiovascular disease. - We studied patients who were acutely admitted with stroke, myocardial infarction and heart failure when their home ward was busy. We found that patients with heart failure had an increased risk of being admitted to another ward and increased mortality. We found weak or no such effect in patients with stroke and myocardial infarction. - Elderly patients are often vulnerable due to ageing and general frailty. We found that mortality or readmission was not affected by the fact that a hospital had an extra busy period. However, hospitals responded with a higher tendency to discharge patients outside daytime working hours. - In a qualitative study, we investigated how health personnel experience and handle capacity pressure in hospitals. The informants reported that the current working day and patient treatment were negatively affected by undercapacity.

Dette prosjektet vil gi bakgrunn og empiri til det nye arbeidet med å utrede helseforetaksmodellen, diskusjonen rundt organisering av akuttkirurgi-kjeden samt samhandling mellom primær- og spesialisttjenesten. Dette arbeidet har dannet grunnlag for betydelig økt helsetjenesteforskning i vår region. I kjølvannet av dette prosjektet har vi også opprettelsen av et registerforskningsmiljø - RegForsk - som er finansiert gjennom en miljøstøtte fra forskningsrådet. Vi har lagt opp til tett tverrfaglig samarbeid med klinisk, epidemiologisk, statistisk og samfunnsvitenskapelig kompetanse. Dette har gitt problemstillinger som er utviklet i tett samarbeid med helsepersonell og andre relevante brukere av våre forskningsresultater. Vi har også erfart økt interesse og pågang fra kliniske miljø for de problemstillingene vi har belyst. Dette forventer vi vil gi bedre og flere forskningsprosjekter med utgangspunkt i tilgjengelige norske registerdata.

In this project we will investigate the impact of system pressures on patient outcomes such as mortality and admissions and readmissions to hospitals. Providing adequate care to patients requires both coordination of care and balancing of capacity between different service levels. Reduced capacity in one part of the system may impair quality and also cause increased pressure in other parts of the system. These issues will be addressed in two parallel parts of the project. In Part 1, our aim is to analyse possible negative patient consequences of reduced length of stay due to non-medical factors. In Part 2 of the project, we analyse the association between general practitioners workload and clinical management of their list patients and their list patients risk of potentially avoidable hospital admissions. Our approach is methodologically ambitious in that we seek to establish not merely associations but causal links, and novel in that it combines data from different levels (primary and specialist health care) and sources (health survey and administrative data) and link these to outcome data such as cause specific mortality. The causal links following patient paths are difficult to assess, as the clinical indications for different paths are likely to be strongly associated with patient outcomes. In our project, we have two targeted approaches to avoid confounding by indication. Firstly we will use a case-only design in order to eliminate time-invariant confounding. That is, we will compare hospital wards and general practitioners with themselves, as the degree of for example capacity pressures varies over time. Next, we will apply an instrumental variable approach, which under certain assumptions has proven to be a fruitful way to assess the causal nature of potentially confounded associations. The project will employ two PhD candidates and two Post Docs.