Skip to main content

Several factors influenced general practitioner participation in the implementation of a disease management programme

Anette Riisgaard Ribe1, Morten Fenger-Grøn1, Peter Vedsted1, Flemming Bro1, Lone Kærsvang2 & Mogens Vestergaard1, 3,

1. sep. 2014
15 min.

Faktaboks

Fakta

Knowledge and awareness about the complex processes involved in implementing changes are the first important steps that must be taken to successfully implement research into evidence-based practice [1]. Disease management programmes in general practice require a high degree of participation to achieve the anticipated impact. Only few studies have quantified factors influencing the general practitioner’s (GP’s) participation in studies [2-4] or in specific disease management programmes introduced by healthcare authorities [5] although such information is a critical foundation for the planning of effective implementation strategies in future [1, 6].

In 2010, the Central Denmark Region invited all 838 GPs based in the 392 general practices located in the region (1.2 mill inhabitants, corresponding to approximately 20% of the Danish population) to participate in a quality improvement project coined the Chronic Care Compass (CCC). The main purpose of the CCC was to improve the quality of chronic care management in general practice for persons with chronic diseases by using various methods including a modified version of the Breakthrough Model [7]. Three chronic diseases were chosen as focus areas for the CCC; diabetes, chronic obstructive pulmonary disease (COPD) and acute coronary syndrome (ACS).

Independently of the CCC, the Central Denmark Region had developed disease management programmes for diabetes, COPD and ACS in collaboration with general practice, municipalities and hospital specialists. These disease management programmes were used in the CCC. The project ran from May 2010 to May 2012.

The Chronic Care Database was established through linkage of administrative registries and clinical databases in the Central Denmark Region in order to support the identification and management of chronic diseases in primary care.

In the present study, we aimed to describe the CCC participation rate among GPs while considering factors characterising the GP and the patients, which may affect participation.

MATERIAL AND METHODS

Design and procedures

Identification of the patient population for the CCC

Patients with diabetes were identified from administrative data in the Central Denmark Region using an algorithm, which is in line with a validated algorithm described in detail elsewhere [8, 9]. Patients suffering from COPD were identified using an algorithm designed to identify patients with COPD in primary care settings [10]. Patients with a history of ACS were identified by extracting data on ACS diagnoses from hospital admissions in the Central Denmark Region,in agreement with the Danish version of the 10th revision of the International Classification of Diseases (ICD-10) [11].

Identification of the study population for the present study

The study population consisted of all GPs in the Central Denmark Region with listed patients with chronic diseases. Consequently, general practices in the region were included in the study if they (as per 1 May 2010) had patients on their list with at least one of the following chronic diseases: diabetes, COPD or ACS.

The practice population of patients with chronic diseases was identified from registers based on algorithms. Patients had to be above 35 years of age at age at baseline; and for a minimum of 12 months before baseline they had to be registered with the same general practice, have been living in Denmark and have been diagnosed with at least one of the chronic diseases. In this study, we defined multimorbidity as the co-occurrence of at least two of the three chronic diseases in the same person. A total of 30 practices established during the study period were not included in the present study. The primary outcome for the study was GP participation in the CCC. Participation was voluntary, but the participating general practices were obliged to register chronic diseases, implement and work with disease management programmes for the three chronic diseases, participate in five CCC education meetings and fill out evaluation questionnaires. In return, each clinic received DKK 24,000 (approximately EUR 3,200) for their participation in the project. If the clinic had more than one GP, the clinic received DKK 10,000 (approximately EUR 1,333) for each additional GP. All GPs and staff were obliged to participate in meetings when participating in the CCC.

Data

Data on the practices, GPs and patients were drawn from administrative registries for the period from 1 May 2009 to 1 May 2010.

Information on the GPs, practices and the use of healthcare services in primary care was obtained from the Danish National Health Service Register [12]. The socio-economic burden within the patient population for each practice was estimated by calculation of the Danish Deprivation Index (DADI) [13, 14]. This index takes a value between ten and 80; higher values indicate a higher proportion of deprived patients. Information on socio-economic factors for the eight key variables included in the index was obtained from Statistics Denmark [15]. Data on participation in the CCC were obtained from administrative services provided by the Central Denmark Region.

For patients, all data were linked using the ten-digit Civil Registration Number assigned to all Danish residents at birth or immigration [16]. Patient information on age, gender and registration with a general practice was obtained from the Danish Civil Registration System [16].

Information on the number of consultations, practice gross income and DADI score was collected for the year before the implementation, whereas all other variables were measured at the start of the implementation period.

Statistical analyses

We included all practices in our main analyses, but performed sub-analyses on solo practices. We calculated the unadjusted participation rate within subgroups and the corresponding 95% confidence interval (CI) as given by Jeffrey’s method [17]. In addition, we calculated the adjusted risk differences (RDs) using binomial regression, i.e. a generalised linear model from the Bernoulli family using the identity link function. In the analyses of all practices, absolute risks were adjusted for type of practice, gender and grouped mean age of the GPs as evaluated per practice. In the sub-analyses performed on solo practices only, all absolute risks were adjusted for gender and age group of the GPs.

Trial registration: Not relevant.

RESULTS

Participation for all practices

A total of 271 practices comprising 609 GPs participated in the project, corresponding to a participation rate of 69.1% (Table 1). Overall, participation was 28.9 percentage points (pp) (CI: 14.3; 43.6) lower among GPs who were older than 60 years than among GPs who were younger than 50 years, 32.2 pp (CI: 19.1; 45.2) lower among GPs who provided few compared with many chronic care consultations, 13.7 pp (CI: 1.7; 25.6) lower among GPs who had a low compared with a medium practice gross income, and 16.9 pp (CI:6.1; 27.8) lower among GPs with a patient population with a medium compared with a low degree of deprivation (Table 1).

Participation for solo practices

Overall, 114 (61.3%) solo practices participated in the project. Participation was associated with the same characteristics as those described for the entire group of GPs (Table 2).

Characteristics of the patient population

The patients listed with participating and non-participating GPs were similar in regard to age, gender and prevalence of chronic diseases (Table 3). Overall, 24,913 persons with chronic diseases (comprising 29.9% of the entire patient population) were not included due to GP non-participation, hereof 13,023 with diabetes (comprising 30.0% of the diabetes population), 11,912 with COPD (comprising 29.7% of the COPD population) and 1.912 with ACS (comprising 30.2% of the ACS population) (Table 3).

DISCUSSION

Seven out of ten GPs in the Central Denmark Region participated in the CCC. The participation was lower among GPs who were older than 60 years, had fewer chronic care consultations, had a lower practice gross income and a patient population with a higher degree of deprivation. Similar results were found when evaluating solo practices only. The patients of the GPs who participated in the CCC were comparable with the patients of non-participating GPs in regard to the distributions of chronic diseases, age and gender.

A recent British report concluded that quality improvement initiatives must consider the heterogeneity in general practice [18]. Little is known about factors influencing GP participation in studies [2-4] or in disease management programmes in primary care [5]. Another British study evaluated the GP participation in a prospective design and showed that forgetfulness and time pressure among the GPs were the main factors inhibiting recruitment [2]. A British review showed that time constraints, lack of staff, lack of rewards and recognition, and an insufficiently interesting research question also reduced GP participation in randomised controlled trials [3]. Finally, another British study found that financial incentives, peer pressure and a desire to improve patient care and maintain professional autonomy and pride were factors that motivated GPs to participate in quality improvement programmes for chronic diseases [5]. These studies all focused on barriers to and facilitators of GP participation in order to address causal mechanisms for non-participation, whereas our study focused on describing the characteristics of GPs, practices and patients influencing GP participation in a disease management programme. However, some overlap and consistency exist.

We found that GPs with lower practice gross income were less likely to participate in the CCC, despite remuneration. This finding indicates that financial incentive is an important, but not a decisive motivation for participation, which is consistent with previous findings [18, 19]. On the other hand, this reluctance to participate could also be explained by the fact that low productivity for GPs is associated with characteristics associated with demanding patients (e.g. old age, low socio-economic status, high use of pharmaceuticals) [20]. Denmark has a fee-for-service healthcare system, which rewards short consultations with presentation of few and simple medical problems. This implies that practices with a complex patient population comprising many older patients, patients with multimorbidity or ethnic minorities will tend to have a lower practice gross income despite excessive busyness. Consequently, the association between low gross income and non-participation could, in fact, be explained by an association between non-participation and time constraints, as also stated in previous studies [2, 3].

Furthermore, we showed a lower participation among GPs who serve a patient population with a higher deprivation score, which could also be explained by time constraints. GPs who serve a patient population with a higher degree of deprivation have formerly been shown to have a lower productivity [20].

In addition, the participating GPs were more actively engaged in chronic care, which indicates that GPs already focusing on chronic care management were more likely to participate. These results are in line with the findings in other studies indicating that the desire to improve patient care, professional pride and an interesting research question were key motivating factors for participation [4, 5].

Finally, the association between non-participation and a GP age above 60 years was unsurprising, as older GPs may have less focus on long-term initiatives in their practice.

Our study has several strengths, including a large cohort comprising all GPs and patients in the Central Denmark Region. This allowed for a population-based design with high statistical accuracy. Therefore, bias due to selection of study participants, loss to follow-up and non-response cannot explain our findings. The Danish National Health Service Register is used to remunerate the GPs. The completeness of this national register is therefore considered to be high [12], and data entries in the Danish Civil Registration System are known to be very accurate [16]. Accordingly, we do not expect any serious information bias on this account.

This study evaluates GP participation for a specific implementation strategy used. Consequently, these data do not allow us to suggest what the results would have been if a different implementation strategy for the disease management programmes had been used by the Central Denmark Region.

Failure to successfully implement evidence-based practice has been suggested to involve barriers at different levels of the healthcare sector (e.g. the individual healthcare professional, the patient as well as the organisational_, social, economic and political contexts) [6]. Our study has primarily focused on characterisation of the participating GPs and their practices, and the collected data do not allow us to address barriers at other levels or to present exact explanations for non-participation among individual GPs.

CONCLUSIONS

Recruitment of GPs to the CCC was successful in terms of participation rates, as a relatively high proportion of GPs participated in the implementation of the disease management programmes. Patients of non-participating GPs did not differ from patients enlisted with participating GPs in regard to type of disease or socio-demographic factors. Despite remuneration, GP participation was associated with being more chronic care active, having a patient population with a lower degree of deprivation and having a higher practice gross income. Our findings suggest that financial incentives alone are insufficient to recruit GPs to the disease management programmes. More efforts are needed to secure participation of more GPs in future disease management programmes. Incentives must be explored prior to recruitment for implementation of specific disease management programmes. Thus, this study indicates that tailored recruitment strategies are critical for ensuring GP participation in future quality development projects.

Correspondence: Anette Riisgaard Ribe, Research Unit for General Practice, Department of Public Health, Faculty of Health, Aarhus University, Denmark. E-mail: ar@feap.dk

Accepted: 25 June 2014

Conflicts of interest:Disclosure forms provided by the authors are available with the full text of this article at www.danmedj.dk

Referencer

REFERENCES

  1. Kitson A, Harvey G, McCormack B. Enabling the implementation of evidence based practice: a conceptual framework. Qual Health Care 1998;7:149-58.

  2. Peto V, Coulter A, Bond A. Factors affecting general practitioners’ recruitment of patients into a prospective study. Fam Pract 1993;10:207-11.

  3. Ross S, Grant A, Counsell C et al. Barriers to participation in randomised controlled trials: a systematic review. J Clin Epidemiol 1999;52:1143-56.

  4. Rendell JM, Merritt RD, Geddes JR. Incentives and disincentives to participation by clinicians in randomised controlled trials. Cochrane Database Syst Rev 2007;2:MR000021.

  5. Spooner A, Chapple A, Roland M. What makes British general practitioners take part in a quality improvement scheme? J Health Serv Res Policy 2001;6:145-50.

  6. Grol R, Wensing M. What drives change? Barriers to and incentives for achieving evidence-based practice. Med J Aust 2004;180:S57-S60.

  7. IHI’s Collaborative Model for Achieving Breakthrough Improvement. Diab Spect 2004;2:97-101.

  8. Kristensen JK, Drivsholm TB, Carstensen B et al. Validation of methods to identify known diabetes on the basis of health registers. Ugeskr Læger 2007;169:1687-92.

  9. Kristensen JK, Sandbaek A, Lassen JF et al. Use and validation of public data files for identification of the diabetic population in a Danish county. Dan Med Bull 2001;48:33-7.

  10. Smidth M, Sokolowski I, Kaersvang L et al. Developing an algorithm to identify people with chronic obstructive pulmonary disease (COPD) using administrative data. BMC Med Inform Decis Mak 2012;12:38.

  11. World Health Organization. International Statistical Classification of Diseases and Related Health Problems (ICD10). Geneva: WHO, 1992.

  12. Andersen JS, Olivarius NF, Krasnik A. The Danish National Health Service Register. Scand J Pub Health 2011;39:34-7.

  13. Vedsted P, Sørensen TH. Praksisscore på DADI baseret på praksispopulationer i 2006-model I og model II. Copenhagen; The Danish GP Organisation (PLO), 2009.

  14. Olsen KR, Gyrd-Hansen D, Sorensen TH et al. Organisational determinants of production and efficiency in general practice: a population-based study. Eur J Health Econ 2013;14:267-76.

  15. Statistics Denmark Press. Statistics Denmark. IDA – en integreret database for arbejdsmarkedsforskning. Copenhagen: Statistics Denmark, 1991.

  16. Pedersen CB. The Danish Civil Registration System. Scand J Public Health 2011;39:22-5.

  17. Lawrence D, Brown T, Tony C et al. Interval estimation for a binomial proportion. Stat Sci 2001;16:101-33.

  18. King’s Fund. Improving the quality of care in general practice. Independent inquiry into the quality of care in general practice in England. London: King’s Fund, 2011.

  19. Glasziou PP, Buchan H, Del MC et al. When financial incentives do more good than harm: a checklist. BMJ 2012;345:e5047.

  20. Olsen KR, Sorensen TH, Gyrd-Hansen D. Association between productivity, list size, patient and practice characteristics in general practice. Ugeskr Læger 2010;172:1192-6.