Background: Despite longstanding experimental evidence about effectiveness of academic detailing, transitioning this research-based concept into acceptable primary care quality improvement services has been slow in the USA. The purpose of this study was to describe primary care uptake, retention and response to an academic-detailing-led service in the USA. Exploration of accessible performance indicators of service acceptance, and feasibility of use of an Australian academic detailing service model were secondary objectives.
Methods: Over a 29-month period, an academic-detailing-led drug and therapeutics information service was offered to all primary care physicians providing ongoing patient care in Fayette County, Kentucky. Two programmes (on type 2 diabetes management and chronic pain management) incorporating up to four office visits were offered.
Results: 102 of 130 (78%) eligible primary care physicians participated in the service, 72% receiving visits for the type 2 diabetes management programme, and 58% the chronic non-malignant pain programme. At all successive encounters, participants expressed a willingness to continue to receive visits. Difficulties were experienced in obtaining appointments for subsequent visits, although on direct enquiry, only one participant explicitly declined further visits. No notable differences existed between physicians accepting visits and those who did not. Across successive visits, passive indicators of satisfaction with the service included: duration of visits, office waiting times, retention of printed materials from one visit to the next, whether physicians wished their extender colleagues to also receive visits, and observed levels of interest and participation within encounters.
Conclusions: Ongoing primary care quality improvement services spearheaded by academic detailing can be acceptable to US primary care physicians in practice.
Statistics from Altmetric.com
In the USA as elsewhere in the world, concerns have been expressed about the slow rate of adoption of safety and quality-improving strategies in healthcare.1 Of additional concern has been the influence of extraneous forces which fashion knowledge for healthcare workers slanted so that patient welfare is overshadowed by other interests.2
Twenty-five years have passed since the original experimental work was performed investigating the power of brief one-to-one interactions between primary care practitioners (PCPs) and trained office-visitors to alter patterns of clinical practice.3 The authors of this pioneering work coined the term “academic detailing” to describe the processes they used in their intervention.
Many other investigators since that time have adopted and adapted the general principles of academic detailing,4 systematically evaluating its effectiveness in different aspects of clinical care. Social marketing, adult learning and diffusion of innovations theory have been described as providing the theoretical basis for academic detailing.5 Despite loosely applied definitions for the academic detailing process, extensive experimental research into its general effectiveness has mostly shown modest to moderate positive outcomes.5–10
Twenty-six systematic reviews of the large body of experimental work on clinical practice improvement strategies in many different settings have recently been summarised.11 It has been concluded that interactive techniques such as academic detailing are the most effective at changing practice as well as patient outcomes.
Economic evaluation of randomised controlled trials of the effectiveness of academic detailing have been less commonly performed. However, in prescribing1213 and cancer care improvement programmes,14 academic detailing has been found to produce economically worthwhile results. Of note, when the inherent value of specific practice changes has been accounted for in these evaluations of academic detailing, even relatively small absolute effect sizes (eg, 5%) have been found to be economically sound.13
Transitioning knowledge gained from all of this experimental research into public health improvement programmes is an important goal for health policy administrators. In Australia,1516 New Zealand,17 Canada,18 the UK19 and The Netherlands,20 public interest programmes have been operating using academic detailing. These programmes have generally used academic detailing techniques as a spearhead for multiple interventions to improve clinical care, its safety, quality and value. In the USA, large health insurers21 have been adapting the technique, and more recently, Medicare Quality Improvement Organisations have been encouraged to adopt social marketing techniques in their interactions with healthcare providers.22
Unless characterised as part of a clinical trial, there have been relatively few publications2324 flowing from this translational work because it has generally fallen outside the scope of scholarly activity. However, as has recently been pointed out,25 publication of information about initiatives of this kind is crucial if healthcare quality and safety is to be advanced, and generalisable knowledge is to be gained.
Largely missing from the literature are useful performance indicators to evaluate operational success of such services. The extent to which such programmes might be accepted and valued by the generality of physicians and their service extenders is a further significant issue. This information from real-life settings is now needed if benefits to primary care practice and public health are finally to be delivered as a result of the experimentation in this field over the past quarter of a century.
The purpose of this study was to describe primary care uptake, retention and response to an academic-detailing-led service in the USA. Exploration of accessible performance indicators of service acceptance, and feasibility of use of an Australian academic detailing service model were secondary objectives.
The geographically delineated area of Fayette County in central Kentucky was selected for service delivery. This area has a regional metropolitan demography reflecting a considerable proportion of middle America (2000 Census figures for Fayette County, Kentucky: Population 260 512, median household income US$39 813, 19% non-white, 13% below the poverty line, one city (Lexington), five rural communities and five historically black rural hamlets). The experience and evolved academic detailing methods of the Australian Drug and Therapeutics Information Service (DATIS) were adopted in this implementation.15
At the outset, a medically led Advisory Board was established to ensure local medical ownership and engagement with the programme. The programme was characterised as a service to primary care physicians, designed to deliver useful evidence-based therapeutic information and advice, independent of commercial or third-party payer interests. The Board chose specific clinical topics of local interest and concern for focused attention in two consecutive office visits spaced apart across time.
Two pharmacist visitors with sound clinical background and training were selected to provide the academic detailing services. The initial topic delivery commenced in February 2003 and dealt with primary care management issues in type 2 diabetes. A second topic concerning management of chronic non-malignant pain was chosen by the Board for delivery from March 2004 onwards. Operations of the programme extended until July 2005.
In addition to the visiting service, PCPs were invited to use the skills and institutional connections of visitors for advice on therapeutic issues at anytime, either during or between visits. On request, continuing medical education credits were also provided for participation with the service.
No charges were levied for services, although PCPs often gave up income-generating practice time to receive visits. An alternative was to receive visits at the beginning or end of the consulting day, or during lunch breaks. When requested, catering was provided for lunch-time meetings: in Fayette County, Kentucky, commercial detailers often provide this form of hospitality in return for PCP time.
Service delivery was offered initially to all non-institutionally based primary care physicians in the County who provided ongoing care to patients with type 2 diabetes (“in-scope” physicians). Then, at the invitation of each physician visited, the service was offered to their associated primary care extenders, (physician assistants and nurse practitioners).
Acceptability of the service was gauged by both the proportion of “in-scope” physicians who agreed to receive visits as well as their subsequent willingness to accept further visits. Performance indicators for the service included features of successive encounters (eg, durations, retention of proffered materials from visit to visit, levels of apparent engagement in the encounter, etc). Uptake, retention and operational performance indicators provide a composite picture of PCP satisfaction with the service.
Staff and material preparation
Intensive training in social marketing and other persuasive communication techniques was provided for visitors using methods described elsewhere.26
Before commencing visits, thorough grounding was provided for visitors in the clinical topic areas. This was achieved through personal review of an extensive bibliography of relevant literature. A comprehensive well-indexed primary-care-focused summary of the literature was then written jointly by the visitors. This material was reviewed by local clinical experts in the topic area and published in a loose-leaf format (a published “topic manual”). In the case of the type 2 diabetes topic, the review extended over 208 pages with more than 600 cited references, and in the case of pain, 130 pages and 320 cited references. This process gave visitors a deep understanding of the current clinical issues involved within each topic area. The manual represented a tangible expression of the visitor’s knowledge and was designed to be left with the PCP as a handy reference source.
A series of key messages were distilled from this review and publication process. These messages, considered and endorsed by the Advisory Board, included elements designed to augment knowledge and/or change clinical practices (see appendix A)
Prior to commencement of the second round of visits on each topic, an additional smaller publication was produced covering issues in depth which had been found to be consistently part of conversations during initial visits. The second round of visits provided an opportunity for coverage of aspects of the topic of particular interest to the practitioner and allowed subtle reinforcement of elements of key messages delivered at initial visits.
A laminated message-summary card was produced to provide consistency in presentation of key messages. It was graphically attractive and did not itemise key messages in the format shown in appendix A. This card was designed to be left with practitioners as an aide-memoire for issues covered during the encounter.
Early in each visiting sequence, office visits were made to specialists clinically active in the topic areas: viz. endocrinologists and pain specialists. These experiences provided further enrichment of the knowledge base of visitors. They also allowed a gauging of attitudes of local specialist opinion leaders.
Office visit details
Office visits were actively solicited, initially through faxed and mailed correspondence, and later through multiple personal approaches to office staff. Finally, if these strategies failed to gain a scheduled visit, personal approaches were made by Advisory Board members generally perceived as primary care opinion leaders in Fayette County.
The structure of visits was always flexible and designed to meet needs of practitioners. Timing was at the convenience of the practitioner and as close as possible to customary practice hours in the office setting. Key message presentation was not didactic in nature but tailored to specific interests and needs of practitioners. Visitors were trained to elicit and understand individual practitioner’s needs so that when features of key messages were presented, benefits to individual practice could be perceived by the practitioner. Close attention was paid to building credibility and establishing trusting relationships throughout visits and ongoing service delivery. Even-handed acknowledgement of alternative points of view and uncertainty in clinical practice was a hallmark of each encounter.
Comprehensive systematic records of interactions were maintained by visitors. In keeping with the service philosophy of the programme, practitioners were not asked to complete questionnaires or documentation as a result of their interaction with the visitor. An exception to this principle occurred when practitioners requested Continuing Medical Education (CME) points needing documentation before accrediting authorities could issue credits.
As shown in table 1, after 29 months of service operation, 72% and 58% of all primary care physicians practising within the defined area received visits on type 2 diabetes and chronic non-malignant pain management respectively. Overall, 78% of primary care physicians practising within the area received the DATIS visiting service at least once on one or other (or both) topics. In around three-quarters of all physicians receiving a first visit on either topic, a second visit on the same topic was then able to be completed. In none of the visited physician practices were DATIS visits to practise extenders discouraged. Whenever asked, physicians agreed that visits for their physician assistant and nurse practitioner colleagues should occur.
Almost one-third of physicians receiving the service were female, and 7% had received their primary medical training outside the US with 58% having graduated more than 15 years prior to the service’s operation; 39% of primary care physicians receiving visits were qualified Family Practitioners and 58% Internal Medicine Practitioners.
No significant differences between participants and non-participants were found in a range of publicly available demographic characteristics of primary care physicians practising within the study area in June 2005. These factors included gender, zip code of the main practice office, type of practice specialty (family practice, internal medicine or general practice), type of degree (DO or MD), years since graduation from medical school, whether or not the medical school was located within the US, whether or not the graduating medical school was the University of Kentucky (which is based in Fayette County) or the number of years since first licensure in the state of Kentucky.
At the conclusion of each initial encounter, practitioners were specifically asked whether they would be willing to receive further visits: all practitioners indicated their willingness to receive such visits.
However, considerable difficulty was experienced in scheduling and then completing subsequent visits as reflected in visits data in table 1. Difficulties paralleled those experienced with obtaining the initial visit. They were primarily associated with persuading office staff of the physician’s stated intentions for further visits. After approximately 2 years of operation, in the small number of cases where it had proven difficult to obtain further visits, investigations were undertaken to establish with the physician in question why further visits had been difficult to arrange. As a result of this survey, a single physician confirmed that they did not wish to continue to receive visits.
Descriptive statistics on the characteristics of visits, and the levels of interest and engagement of PCPs at visits are provided in table 2.
For between half and two-thirds of appointments no waiting time was experienced by visitors. When waiting time was experienced beyond the appointed time, the mean duration was around 15 min.
Initial visits on individual specific topics were characteristically more than 20 min in duration. There was no difference between the duration of second visits on each topic, but a significant increase in duration was observed for the first visit on pain, relative to the duration of the initial visit on type 2 diabetes management.
Most practitioners were seen alone one on one, and catering at visits was needed for about one-third of visits. At roughly half of the diabetes visits, the practitioner indicated a wish to receive CME credits for the visit, but this proportion dropped to one-third for the pain topic.
The level of participation by practitioners during the one-to-one interaction and their interest in the topics presented (as perceived by the visitor) were generally high. Importantly, around 90% of practitioners appeared interested in the topics discussed. More than half were considered to have better than a moderate level of participation in visit discussions, 90% being considered to have either good or high participation.
The level of interest in the service itself was similarly very high, with most participants being keenly interested. At initial visits, there was commonly a curiosity about the nature of the programme, and a predictable level of suspicion as to biases or lack of independence from commercial or third-party-payer interests. However, this misconception was generally easily dispelled, particularly as subsequent visits ensued over time.
Finally, for those practitioners receiving follow-up visits on either topic, in about half of all visits the topic manual left in the office at the previous visit was clearly visible in the office or could easily be found. Informal feedback consistently reinforced the view that the manual proved to be of use. In a quarter to a third of cases, informal discussion at subsequent visits revealed that manuals had indeed been read since the date of the previous encounter.
At effectively all visits, the published manual of clinical information and associated detailing card were left with the practitioner. During discussions, a range of other printed materials relevant to topic key messages were provided. In the case of the pain topic, more than two-thirds of practitioners accepted pads of tear-off pain-diary sheets for patients, agreeing they could be helpful for managing their patients. A similar proportion accepted checklists for prescribing controlled substances and for assessing patients for their addiction potential, as well as a draft patient contract for commencement of controlled substances in pain management.
In the case of the type 2 diabetes management topic, around half of the practitioners visited accepted both patient information cards (59%) and a care-checklist for the practitioner’s case records (47%) acknowledging that they would be helpful for patient care. These materials had been produced by the State-supported Kentucky Diabetes Network (KDN). Prior to commencement of the visiting service, these same materials had been sent by mail to all Kentucky PCPs by the KDN. However, in none of the 122 initial visits on this topic could practitioners recall having seen these practice aids prior to the service visit.
The time spent during each interaction covering the key messages of the respective topics is shown in fig 1.
The time spent during second visits on each topic covered specific aspects of either diabetes or pain management as appeared to match interests and needs of individual practitioners. Overt reiteration of topic key messages during these second visits did not necessarily occur. Nevertheless, key messages were often touched upon in these follow-up visits when different aspects of the topic of specific interest to the individual practitioner came under discussion.
The printed materials used at follow-up visits provided a framework for second-visit interactions. These materials had been published in the same loose-leaf half-letter size as the original manual, and visitors asked practitioners if they could add them to the practitioner’s folder towards the end of the interaction. In this way, a transparent means was used to establish the location of the topic manual which had been left at previous visits.
With respect to the proportionate time spent dealing with various topic key messages at first visits, no notable differences were apparent between the two visitors, nor with respect to years since practitioner graduation or licensure, practitioner gender or specialisation.
After two and a half years of operation of this academic detailing service, almost four-fifths of primary care physicians working in a defined geographic area had taken up the opportunity to participate. Once an initial visit had occurred, effectively none of the physicians said (in response to direct questioning) that they did not wish to continue to receive further visits.
Apart from this expressed willingness to continue to receive services, there are a number of additional indicators of participant satisfaction. In all cases, visited physicians who employed physician extenders in their office practices wished the service to be provided to their PA and ARNP colleagues. Significant periods of time were generally spent in each academic detailing encounter. This time represented lost opportunities which otherwise could have been used by the PCP for regular patient consultations or other office activities. Consistently short waiting times for appointments and generally very positive receptions from physicians were also experienced. Additionally, despite the plethora of printed information delivered to primary care physician offices, in almost half of encounters, the service’s manual of clinical support information delivered at previous visits was easily found. In about a third of follow-up visits, the manual was located on the desk or shelf in the consulting office despite a mean time lapse of roughly 6 months between visits.
Difficulties were experienced in gaining follow-up visits with some participants despite their directly expressed willingness during previous encounters. These difficulties were largely associated with PCP office management conventions, customarily short open-diary times for future appointments and operational factors related to the lack of permanent funding for this service. Given the past performance of the Australian academic detailing service model, with further experience and secure service funding it is expected that many of these difficulties would be overcome. For recruitment of the residual one-fifth of practitioners working in the area, it is anticipated that given time and a range of further topics, substantially complete uptake could be achieved. To gain a crucial initial visit, it is helpful for discussion topics to be of particular interest to previously unrecruited physicians.
Without using cash or other incentives, PCP uptake of service-oriented academic detailing programmes during the 1990s in Australia reached 95% with virtually complete retention.15 In Canada, a longstanding service-oriented programme in North Vancouver also has reported an uptake of more than three-quarters of eligible PCPs.27
In the USA, there have been no reports of recruitment and retention rates for all primary care physicians in an area offered similar long-term continuing education programmes spearheaded by academic detailing relationships.
A short-term academic detailing programme operating in broadly dispersed regions of Pennsylvania in 2003 achieved an uptake of physicians of 31%.28 However, this programme focused on a single topic, and recruitment was only possible during a 5-month period. Academic detailers worked solo and part time in the programme, while the Kentucky programme employed full-time detailers from a central location in one discrete geographic area. Pennsylvania physicians were identified for recruitment based on their care of numbers of elderly beneficiaries enrolled in a state-sponsored drug-assistance programme. These programme features may have contributed to differences observed in recruitment. Five months after commencing the Kentucky programme, roughly 60% of the total possible PCPs had received a first visit.
The Kentucky programme, modelled on Australian experiences, was carefully structured and promoted as a service to PCPs. It also focused on building trusting individual personal relationships which sales theory has long affirmed to be important for achieving success in commercial dealings.29 It is plausible that this service orientation and relationship focus contributed to the programme’s success in recruitment and retention.
While about half of the participants verbally expressed a wish to gain continuing education credits for visits on diabetes management, only a third completed the paperwork necessary to allow this to happen.
Beyond service recruitment and retention, a range of parameters are presented here which can be used across time to gauge success of academic-detailing-led programmes designed to improve quality and safety in healthcare. Apart from retention by the service and success in gaining future appointments, waiting times, duration of interviews, retention of materials from visit to visit and visitor-observed attitudes offer useful gaugings of recipient opinion and receptivity. Such passive observations across time can build a helpful picture of the utility of this form of continuing education service.
Future programmes of this kind need to explore further passive indicators to gauge service performance. Benefits from the use of two consecutive visits on the same topic might be replaced by judicious selection of sequential topics allowing subtle reflection of key message elements from previous encounters (eg, management of diabetes followed by a programme on management of hypertension).
A notable feature of this programme has been the nature of the cohort of physicians it has attempted to serve. Every primary care physician providing ongoing patient care in a specific geographic region was offered the service. At least in so far as the target region reflects primary care practice in other parts of the US, recruitment and service retention outcomes represent encouraging whole-population penetration.
If a very high proportion of PCPs find it worthwhile to continue participation in this type of programme (with benefits proven through 25 years of experimental research), it is reasonable to assume that participants will include those practitioners who pay less attention to their own continuing professional education. From a public health standpoint, it is for this subgroup that this type of programme will potentially be of greatest value.
PCP reports of experiences with academic detailers have not been canvassed. However, where programmes attempt recruitment and retention of all PCPs practising in a geographic area, a significant proportion are unlikely to provide written feedback of their own volition. Biases in opportunistic opinion sampling are well recognised with extremes of opinion likely to be over-represented, limiting generalisability. Additionally, the act of requiring or seeking written opinion is likely to compromise achievement of key outcomes: viz. recruitment, and ongoing participation.
Exploration of clinical practice changes and outcomes across time as reflected in databases of third-party payers are to be reported elsewhere.
The extent of willingness of contemporary US PCPs to accept visiting services from professional academic detailers is an important unresolved issue for administrators of programmes designed to improve quality and safety in primary healthcare.
This programme which operated over more than 2 years in Kentucky has demonstrated that a substantial majority of PCPs are willing to participate in such programmes in regional metropolitan US settings. It also has demonstrated that a range of passive observations can provide helpful indicators of the acceptability of such services.
The encouragement, advice and work of DATIS US Advisory Board members, J Borders, S Matheny, A Tzouanakis, B Casey and R Littrell, is also acknowledged with deep gratitude.
APPENDIX A PROGRAMME KEY MESSAGES
MANAGEMENT OF TYPE 2 DIABETES
1. Identification of people at risk of type 2 diabetes and pre-diabetes (impaired fasting glucose (IFG) or impaired glucose tolerance (IGT)):
prevalence of type 2 diabetes and pre-diabetes (IFG, IGT) in USA and Kentucky;
at-risk groups: emphasis on increased risk of small vessel (microvascular) and large vessel (macrovascular) disease and the influence of lifestyle on risk;
merits and weaknesses of venous plasma glucose, fasting plasma glucose, oral glucose tolerance test (OGTT) and their place in clinical practice for diagnosis of type 2 diabetes and identification of pre-diabetes (IFG and IGT);
lifelong lifestyle modifications are vital: nutrition, exercise, body weight management, smoking cessation.
2. Optimise glycaemic control:
emphasise importance of long-term glycaemic control by monitoring A1C levels 2–4 times a year, with a target of <7% (American Diabetes Association 2003);
emphasise importance of A1C targets as a patient motivational aid;
emphasise that evidence has shown a 1–2% sustained reduction in A1C significantly reduces the risk of developing microvascular complications, slows the progression of microvascular complications and may also reduce the risk of developing macrovascular complications;
metformin is the preferred initial drug therapy of choice where it is not contraindicated: randomised control trials have shown that metformin reduces the risk of any microvascular end point, the risk of a number of macrovascular end points, and only metformin has been shown to reduce macrovascular complications and all-cause mortality in obese patients with type 2 diabetes;
sulfonylureas are often considered first for non-overweight patients but are associated with an increased risk of hypoglycaemia and weight gain.
3. Optimise blood pressure control:
emphasise that evidence has shown tight blood pressure control is more effective than intensive glycaemic control in protecting against microvascular and macrovascular disease.
4. Regular and systematic monitoring/management of all cardiovascular risk factors present:
consider specific cardiovascular risk factors including those of lifestyle, smoking, hypertension, and dyslipidaemia: glycaemic control alone does not eliminate the excess cardiovascular risk of diabetes;
consider use of aspirin;
assessment of microalbuminuria.
5. Individualisation of management:
emphasise importance of individualised monitoring and targets, patient self-management, lifestyle modifications and pharmacotherapy;
any movement toward the desired goal is beneficial: weight loss, blood pressure, lipids, A1C.
MANAGEMENT OF CHRONIC NON-MALIGNANT PAIN
1. Better evaluation and improved documentation of the pain patient:
prevalence of pain in USA and Kentucky;
pain scales; utilise one scale only with each patient;
physical exam, define aetiology of pain;
emphasise importance of documentation;
document everything (initial and follow-up visits, including thoughts, medicines prescribed, doses, refills, etc);
importance of a treatment plan and consent (written or oral);
document treatment goals—decrease pain to tolerable levels and increase function (goals—hobbies, work, etc).
2. Improve discrimination in use of non-opioid therapies:
Acetaminophen first line due to improved safety profile over NSAIDs: WHO step therapy;
similar risk between NSAIDs for CV complications (HTN, renal dysfunction, oedema);
increased risk of CV disease (specifically MI) with COX-2 selective NSAIDs still controversial;
GI toxicity: lack of RCTs proving difference for NSAID versus COX-2 selective agents in low risk group and/or those on concomitant aspirin therapy;
COX-2 inhibitors: RCTs show decreased risk for GI toxicity in high risk groups, but poor evidence for benefit in patients with pre-existing GI complications;
potential benefit of adjunctive and alternative therapies.
3. Screening for addiction:
define addiction, tolerance, physical dependence, pseudoaddiction;
tools that can be used for screening;
4. Better management of opioid side effects:
tolerance developed to most side effects excluding constipation;
utilise laxative/stool softener in those without contraindications;
misconceptions (need for continued dose escalation, addiction causation, etc).
5. Individualisation of therapy:
individualise documentation, using specific words from interview and thoughts, for each patient;
individualise treatment plan utilising combination of pharmacological and non-pharmacological approaches.
Funding: Financial support for this programme was provided by the CDC under the Department of Health and Human Services, Public Health Services Grant no 02054, Grant no E11/CCE421825-01 and RO1 C100188-01.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.