Inductive analysis
The inductive analysis revealed eight types of management strategies that were perceived as influencing LVC use. Five of these were management strategies at the regional level: scorecards, clinical decision support, lectures, quality assurance systems and financial systems. Three strategies, process strategies, locally held lectures and discussions about guidelines, was decided at the center level.
Score cards
Most centers used score cards to track their own performance. Some respondents believed this had influenced LVC, whereas others did not think so. Some of the key performance measures were decided regionally by those financing health care, others were decided by the regional management teams, and finally, some of the measures were added by the centers themselves. Most key measures had a general goal of improving care – e.g., patient safety – but some were described as more specifically related to LVC; e.g., number of antibiotic prescriptions and cost of lab tests per patient visit.
We use it to follow up guidelines, routines, and other templates. So, I believe you can say that it has an influence. Sure, we haven’t had a specific focus on it (reducing LVC) or set a clear agenda for it. But at least indirectly. (IP4)
Guidelines
Guidelines was also a possible way to influence LVC use, with differences between centers concerning which guideline was used. The guidelines were described in terms of written instructions on how to manage different types of diagnoses within the health care system. Region Stockholm administrators developed their own guidelines based on a combination of local knowledge and national guidelines. The regional guidelines were published on a website (viss.nu) and included both guidelines on how to manage different types of diseases and how the responsibility should be divided between primary care and specialist care. Participants from the other two counties did not describe similar regional guidelines but used either a national decision support system called Internet Medicine or used the guidelines developed by Region Stockholm. Like scorecards, the guidelines had a general goal of improving care, even though some of the guidelines were possible to relate to LVC de-implementation.
The centers differed in their perceptions on whether the guidelines influenced LVC use. Some described that the guidelines being written by specialists and designed for specific diseases led to more LVC: when a patient wanted treatment for a symptom that could be due to several diseases, this meant that if they were to follow all the guidelines it would lead to unnecessary lab tests, examinations, and medications:
But in many of the groups, there is no primary care physician and only other specialists; and that leads to more test ordering and examinations, since they are used to another patient population. (IP3)
Others perceived the guidelines as leading to less LVC, as they often recommended a stepwise approach to lab tests and examinations and encouraged the physicians to limit their use of certain treatments.
You do not order the full set of tests at once, and that by definition means that you will order fewer lab tests – that you will order the right lab tests, and that you will find the right track for your assessment (of the patient) sooner. (IP1)
Still others perceived that the guidelines had no impact on their use of LVC because none of the guidelines were written specifically to target LVC:
It doesn’t matter how many decision supports systems you have – there are so many of them already – but if you do not trust yourself and your own assessment, you will not proceed with the patient and a lot of unnecessary things will be done. (IP10)
Lectures
Education in the form of lectures –at the regional level, targeting all physicians, was perceived by some as influencing LVC use. The perceived influence of these lectures differed among the managers. Some perceived the regional lectures as influencing the use of LVC because they were often based on clinical decision support, which describes both what should be done and what should not be done.
The theme was use imaging correctly and appropriately. It was held by a specialist in radiology and covered both when you should order an x-ray and when you should not. (IP8)
Others perceived that the lectures had no LVC-relevant content at all:
It has not been an explicit theme. We focus on diagnoses; we rarely focus on LVC. (IP3)
Quality assurance systems
Several quality assurance systems were described in the interviews as strategies that were or could be used to reduce LVC. All centers had access to the systems and could review their data as often as they liked to improve their performance. Examples of data included how many patients they had with different diagnoses, how many prescriptions of certain medications were ordered, and how many of a specific lab test had been ordered. They also had access to benchmark data on the different results, making it possible to compare their results with other centers both in the county and nationwide. Most performance measurements were related to a general goal of improving care. Managers differed in how often they reviewed their own data and how they used it. One common method was to review all data once per year, choose one measurement that they wanted to improve and include it in the scorecard. Based on the chosen results, center staff could also make a performance improvement plan. However, the work related to the quality assurance systems rarely focused on LVC. Still, the quality assurance systems involved many different variables so that it would be possible to use the system to support reduced use of LVC.
You can pick an example and compare your data to a benchmark for all health care centers. I believe that it is a very successful strategy for a discussion. Like if we look at the lab test sedimentation rate, and then find that wow, we really order so many more of those tests than what is normal for all centers. (IP7)
Financial system
The financial system also had the potential to influence LVC use. Many of the managers described that the previous system, under which the centers were mainly paid for patients’ visits to a physician, had steered the centers toward always booking the patients with physicians when they contacted the center. This was described as increasing the use of LVC, since the physicians’ main tools for managing patients’ symptoms are lab tests, examinations, and medications.
When they reformed the system, we got paid so much more for visits to the physician than a visit to the nurse. This worked as an incentive to make all patients visit the physician, even for things that were not so important. When the patient met the physician, a throat test was ordered even though it was not needed, or maybe antibiotics were prescribed even though the patient did not need them. (IP1)
When the financial system changed to mainly reimburse the centers for the number of listed patients, the incentive to schedule all patients with a physician decreased. However, this system had its disadvantages. Reimbursing the centers for the number of listed patients led to an increased emphasis on having satisfied patients who remained listed at the center. Some managers expressed a concern that this created an incentive for physicians to act in accordance with patient requests for specific examinations or medications to reduce the risk of losing them as listed patients.
When patients are expected to list themselves at the center, it implies that we should be available and accommodating, and there is a slight risk that this will lead to a drift (in following guidelines), especially when it concerns something that is harmless for the patient: to order that lab test, conduct that examination or prescribe that medication. (IP6)
Another LVC-relevant aspect of the financial system was the monitoring of examination and lab test costs. This monitoring encouraged the centers to try to reduce them by discussing them on a regular basis. This surveillance could potentially help reduce the use of lab tests in general, including LVC lab tests.
…and of course, lab tests, since it is an expensive cost for the center, this is something that is being looked at. (IP2)
Center-specific strategies
Three strategies were found on a center level. Process strategies, locally held lectures and discussions about guidelines. Process strategies included different ways of changing work processes to influence the way each center functioned. These strategies were developed at each center by the manager in collaboration with the personnel. The examples described by the informants were often directly aimed at reducing LVC. Some of the process strategies were influenced by regional strategies (i.e., scorecards, decision support and financial system), but some were designed based on initiatives from the managers or center personnel. One example of a process strategy was to stop scheduling patients with a physician when other professional categories were better options. It was argued that scheduling them to see another profession such as a nurse or a physiotherapist reduced the likelihood of the patient receiving LVC, as the other professions did not have access to ordering tests and examinations or prescribing medications.
We try to work a lot with who the patient should meet first and try to avoid it being the physician, since they are so good at ordering lab tests and prescribing medications or sick leave, and many things that are not in the patients’ best interest. (IP6)
They also had different processes for changing routines to reduce the use of specific LVC practices. Examples of routines included restricting what examinations junior physicians could order without approval from a senior physician, and changes in standard test ordering forms that created hurdles for the physicians to order examinations and tests that were considered of low value.
The normal lab tests that we use at the center are grouped in the ordering system to make it more efficient. For this to work, we need to have a couple of people who review these grouped tests and remove those that are no longer needed. (IP2)
A more general strategy to reduce or avoid LVC was to work toward continuity in contact between the physician and the patients by making sure that the center had a good work environment. This would reduce staff turnover and make it easy to recruit highly skilled professionals.
I can tell you right away that the most important thing is that you have a physician who is a specialist within primary care, who knows his or her patients and who is experienced in examining patients and can trust their own findings. That is what it is all about. That means education, experience, and continuity in relation to patient contacts. (IP10)
Locally held lectures included sharing of knowledge related to LVC that individuals had acquired and invited experts to present at their meeting. These were also perceived by some as influencing the use of LVC. Some of the lectures were influenced by the regional strategies (lectures and guidelines) whereas others were planned entirely based on local initiatives.
Discussions about guidelines was the final local strategy that was center specific. Several of the centers described holding regular meetings within the entire personnel group or among the physician group to discuss how to manage different diagnoses. Those discussions included what they should do and what they should not do (i.e., LVC issues).
We also have a lot of discussions within our center on how we manage (different patient symptoms), what we should manage, and we use each other for help. (IP9)
Some of the discussions were based on the personnel’s own clinical experience whereas other were in relation to guidelines.
When we have meetings, we discuss how to manage different patient cases and look at the guidelines.
Deductive analysis of the mechanisms involved in the strategies
To understand the mechanisms involved in the strategies, we analyzed the above LVC-related strategies in terms of the three-term contingency and rule governing.
Three-term contingency
None of the five strategies decided on a regional level could be interpreted as influencing antecedents or consequences in the physicians’ environments directly. The only way for these strategies to change the direct contingencies would be via changes at the centers. For instance, feedback from the scorecards on costs related to lab tests did not directly impact the physicians’ ordering of these tests but encouraged centers to develop their own strategies to reduce costs. These could then impact physicians in turn. The same type of process applied to all management strategies developed at a regional level, including guidelines and education.
Contrary to the regionally developed ones, the strategies developed by the centers could be interpreted as influencing direct contingencies for the physicians. For instance, removing LVC lab tests from standard ordering sets and not permitting junior physicians to sign their own imaging orders made it difficult to use these practices. These changes were similar to behavior analysis principle of increased response effort, which increases the amount of effort needed to perform a behavior related to a reinforcing consequence. Scheduling patients with professions other than physicians removed the situation (i.e., the patient interaction), and thus the contingencies. Planning for continuity in contact between the physician and the patients could reduce the probability for adverse reactions from the patients when saying no to prescribing LVC. All strategies developed by the centers could be interpreted as aiming to reduce behaviors related to LVC use rather than increasing behaviors related to not using LVC.
Rule-governing behavior
Three strategies could be interpreted as trying to influence behavior through rule governing. The regionally developed strategies, lectures, and guidelines, as well as center-specific lectures and discussions on how to reduce the use of specific LVC practices, all included instructions on what not to do with an implied consequence to improve patient health. However, this consequence was not always sufficiently immediate for a physician to experience (i.e., the effects of not using LVC are not always noticeable from a short-term perspective). The lack of immediate consequences for following the rule, such as experienced health improvement for the patients, decreases the probability of guidelines and education acting as rules governing the behavior. The locally held lectures and discussions about guidelines provided socially mediated consequences for following the rules and can thus be interpreted as having a stronger potential for functioning as a rule.
[Figure 1]
Figure 1. Strategies divided based on where they were initiated (regional or center level) and by their function on influencing behavior (antecedents and consequences or rule-governed behaviors).
All strategies are illustrated in Fig. 1 and are divided both based on which level of the organization each strategy was initiated (regional or center level) and in what way they influenced LVC-related behaviors (antecedents and consequences or rule-governed behavior).