Guidelines should consider clinicians’ time needed to treat

Squeezed time

Several studies have shown the impossibility of meeting all guideline recommendations. For example, a simulation study applying all guidelines for preventive care, chronic disease care, and acute care to a panel of 2500 adults representative of the US population estimated that US primary care physicians would require up to 27 hours a working day to implement (and document) all applicable guidelines.2 To fully satisfy only the recommendations from the US Preventive Services Task Force would require 7.4 hours a day.3

Similarly, to implement the European hypertension guidelines in Norwegian adults, Norway would need more general practitioners than are currently in practice.4 And in the UK, implementing all lifestyle interventions recommended by the National Institute for Health and Care Excellence (NICE) may require more physicians (from all specialties) and more nurses than currently available, according to our estimates (unpublished data). Furthermore, healthcare policies also need to account for the time clinicians should spend listening in silence, noticing carefully, and co-creating sensible plans of care with patients.5 If clinicians followed all pertinent guidelines, there would be no time left to care for other ill patients and the healthcare system would collapse.

Clinicians clearly have to make choices about which recommendations to follow in which patients. This involves interpreting, prioritising, and applying them to a diverse group of patients with needs and wishes that may or may not align with practice guidelines. But without guidance, and under the pressure of time, decisions about which guidelines to implement at the point of care will be implicit, variable, and potentially misguided.

To help guide such prioritisation, policy makers may choose a few of the guideline recommendations and tie them to incentives, such as publicly reported quality metrics or pay-for-performance schemes. However, since clinician time is finite and scarce, the chosen interventions and patient groups will inevitably be prioritised at the expense of other interventions and other patient groups.67 Thus, time spent implementing a particular guideline may carry a substantial opportunity cost, and the element of clinical care that is lost might be of more benefit than what is gained.

One strategy to address this problem would be for guideline panels to estimate the time needed to implement an intervention when determining the direction and strength of recommendations. A recent review of all lifestyle interventions recommended by NICE found that clinician time was not considered or estimated for any of the recommendations,8 and the US Preventive Services Task Force does not routinely estimate clinician time needed to implement its recommendations.9 Clinician time is not explicitly included in GRADE’s (Grading of Recommendations Assessment, Development, and Evaluation) widely endorsed and used evidence-to-decision frameworks for guideline development.10

Accounting for real world implementation of guidelines would require guideline bodies to account for time constraints, estimate them, and consider clinician time in a structured way. To this end, we propose that guideline setting bodies explicitly consider the “time needed to treat.”

Time needed to treat

The time needed to treat (TNT) makes explicit the estimated clinician time needed to improve the outcome for one person in the targeted population (TNT_NNT), the clinician time needed to implement the recommendation for all eligible people in a practice (absolute TNT), and the proportion of total clinician time available for patient care that would be needed to implement the recommendation for all eligible patients (relative TNT). Box 1 lists the main factors affecting TNT estimates.

To illustrate the use of TNT, we consider how it might apply to the NICE guideline “Physical activity: brief advice for adults in primary care.”11 This guideline recommends general practitioners (GPs) use questionnaires to screen all adults for physical inactivity and record the outcomes. For those not meeting the recommendations on physical activity, the clinician should give “brief advice” tailored to the individual, discussing motivation and goals; current level of activity and ability; circumstances, preferences, and barriers to being physically active; and health status. The brief advice should contain information about local opportunities to be physically active for people with a range of disabilities, preferences, and needs. The guideline encourages a written outline of the advice and goals, recording the outcomes of the discussion, and follow up in subsequent appointments.11

Studies cited by the guideline suggest that offering this advice would take at least 10 minutes a patient.12 All adults from age 19 are eligible for screening, and roughly 40% of screened adults (the proportion of adults who do not meet the UK recommendations on physical activity13) would then be eligible for brief advice.

To estimate TNT_NNT, we need to know the number needed to treat for the recommended intervention. Based on the NICE evidence review underpinning the guideline,12 14 people with a sedentary lifestyle need to get brief advice on physical activity for one more person to report an increase in physical activity (the review found no significant beneficial effect on cardiorespiratory fitness, mental health, or other outcomes). Given that clinicians need to screen 35 people (which takes a minute per person or 35 minutes) to identify 14 people with a sedentary lifestyle, and then spend 10 minutes with each one giving brief advice (total of 140 minutes), it takes 175 minutes or 3 hours of GP time for one more person to increase their self-reported physical activity.

To implement this recommendation in a general practice of 2000 adults (the average population per full time GP in England14), the absolute TNT can be estimated as follows: it would take 1 minute per person to screen all 2000 adults eligible for screening, for a total of 2000 minutes. Eight hundred adults (40% of 2000) would be eligible for brief advice, which at 10 minutes per person would require a total of 8000 minutes. The absolute TNT in a practice of 2000 adults is thus 10 000 minutes, or 167 hours, of GP time a year.

The practice TNT can also be expressed in relative terms by dividing the absolute TNT by the total time available for direct patient care per GP per year. A GP who spends 60% of their 40 hours a week in face-to-face patient care and who works 47 weeks a year in the practice will have a total time for direct patient care of 1128 hours. The 167 hours needed to implement the NICE physical activity guideline will thus represent 15% (167 of 1128 hours) of the GP’s yearly total face-to-face time with patients.

Including time in guidelines

Alongside other relevant factors, guideline panels should consider TNT estimates when deciding whether and how strongly to recommend an intervention. To help guideline panels do this, guideline development frameworks should set out how to account for clinician time for implementation.

GRADE’s evidence-to-decision frameworks are widely endorsed and used to help guideline panels adequately and transparently consider all relevant aspects when issuing recommendations.10 The frameworks explicitly encourage panels to consider the following criteria: the priority of the problem, certainty of the evidence, benefits and harms, outcome importance, balance of benefits and harms, resource use, equity, acceptability, and feasibility, with a set of guiding questions for each criterion.10 Time to implement fits well in considerations of feasibility and acceptability. When considering these two criteria, panels could review the time needed to treat associated with their recommendations. They can further consider, when accounting for other ways clinicians could potentially spend their time, if spending the required proportion of available clinician time on the recommended intervention is likely to be acceptable and feasible to key stakeholders. Estimating the TNT for other clinical guidelines could facilitate this assessment of opportunity costs.

How TNT could affect decisions

Explicitly considering clinician time by estimating TNT could turn a “strong recommendation in favour” into a weak one or even turn a “weak recommendation in favour” into a “weak recommendation against.” TNTs may prove particularly useful when guidelines make recommendations that apply to a large fraction of the general population and when the beneficial effects of the recommended intervention are, in absolute terms, small. For example, for the NICE physical activity guideline outlined above,11 estimating TNT would have, or should have, given pause to the panel advising that GPs spend 15% of their total time for direct patient care delivering an intervention without any evidence of improvement in long term morbidity and mortality.12 In other words, the high relative TNT may have warranted downgrading the recommendation from “strong in favour” to “weak in favour,” or even to “weak against.”

TNT estimates may also prompt guideline panels to adapt recommendations to reduce their footprint on clinicians’ time. High TNTs, for example, could lead panels to simplify the intervention or target a narrower group of eligible people (ie, set a higher threshold for eligibility). The high cost in clinician time of the recommendations on hypertension care in Norway,4 for instance, could be reduced by restricting the interventions to high risk populations (raising the threshold for treatment) or by reducing the frequency of blood pressure measurement. Other ways to reduce the time required for clinicians include delegating tasks to other healthcare staff or promoting the implementation of technological approaches.15 However, delegation may reduce TNT for one group of clinicians but overwhelm another, and few if any technological applications actually save clinicians’ time. The opportunity cost problem is therefore likely to persist.3

In settings with too few clinicians even the most valuable recommendations may lead to high relative TNTs. Making explicit the cost in clinician time for each recommendation may help to prioritise actions to be taken in the limited time available. In such situations, an alternative solution to high relative TNTs may be expanding the clinical workforce and distributing limited resources more equitably.

Estimations of the clinician time needed to implement a practice guideline are likely to focus the scope and reduce the number and strength of recommendations that panels issue, increase the credibility of the remaining recommendations, and thus increase the likelihood of clinicians implementing the most valuable recommendations. This may also leave more time and capacity for clinicians to make care plans that fit each patient, particularly those living with biological complexity (eg, multimorbidity) or socioeconomic complexity (eg, material and social deprivation).16

Realising the benefits

Because TNT estimates do not directly incorporate important aspects such as the relevance or certainty of the beneficial effects, the same TNT may be acceptable for one recommendation but not for another. The ultimate goals of estimating TNT are to ensure that clinicians and patients spend their limited time together on what is most important for the individual patient and to improve access to care for patients with the greatest care needs.

Guideline panels also have limited time available, and we recognise the importance of not adding unnecessary burden to the already challenging processes of guideline development. Although estimating TNT will add time, one of the strengths of the method is its simplicity, meaning that its consideration will not require advanced methodological skills.

After scrutiny and use, TNT may need revision and improvement. Users may consider more sophisticated models of healthcare utilisation and measures of efficacy to finetune the concept. There are, however, merits of keeping the estimation of TNT simple and transparent and its interpretation accessible for policy makers, clinicians, and patients. Researchers will need to develop methods by which TNT estimates can be used to prioritise recommendations across guidelines that apply to the care of a particular population.

In the meantime, guideline authors should routinely consider the time taken to deliver relatively low yield interventions applied to large proportions of the population. Doing so will appropriately diminish enthusiasm for such interventions, moving strong to weak recommendations and recommendations in favour to those against. The net result will relieve clinicians of impossible sets of recommendations and increase the net benefit of their care.