Comparison of out of hours care provided by patients' own general practitioners and commercial deputising services: a randomised controlled trial. I: The process of careBMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7075.187 (Published 18 January 1997) Cite this as: BMJ 1997;314:187
- D K Cragg, lecturera,
- R K McKinley, senior lecturerb,
- M O Roland, professor and director of research and developmentc,
- S M Campbell, research associatea,
- F Van, research associatea,
- A M Hastings, lecturerb,
- D P French, research associateb,
- T K Manku-Scott, research associateb,
- C Roberts, senior research fellowc
- aDepartment of General Practice, University of Manchester, Rusholme Health Centre, Manchester M14 5NP
- b Department of General Practice and Primary Health Care, University of Leicester, Leicester General Hospital, Leicester LE5 4PW
- cNational Primary Care Research and Development Centre, Williamson Building, University of Manchester, Manchester M13 9PL
- Correspondence to: Dr McKinley
- Accepted 21 November 1996
Objectives: To compare the process of out of hours care provided by general practitioners from patients' own practices and by commercial deputising services.
Design: Randomised controlled trial.
Setting: Four urban areas in Manchester, Salford, Stockport, and Leicester.
Subjects: 2152 patients who requested out of hours care, and 49 practice doctors and 183 deputising doctors (61% local principals) who responded to those requests.
Main outcome measures: Response to call, time to visit, prescribing, and hospital admissions.
Results: 1046 calls were dealt with by practice doctors and 1106 by deputising doctors. Practice doctors were more likely to give telephone advice (20.2% v 0.72% of calls) and to visit more quickly than deputising doctors (median delay 35 minutes v 52 minutes). Practice doctors were less likely than deputising doctors to issue a prescription (56.1% v 63.2% of patients) or to prescribe an antibiotic (43.7% v 61.3% of prescriptions issued) and more likely to prescribe generic drugs (58.4% v 32.1% of drugs prescribed), cheaper drugs (mean cost per prescription £3.28 v £5.04), and drugs in a predefined out of hours formulary (49.8% v 41.1% of drugs prescribed). There was no significant difference in the number of hospital admissions.
Conclusions: By contrast with practice doctors, deputising doctors providing out of hours care less readily give telephone advice, take longer to visit at home, and have patterns of prescribing that may be less discriminating.
In response to a request for out of hours care, practice doctors are more likely than deputising services to offer telephone advice
When patients are visited at home, practice doctors can get there sooner
Practice doctors give fewer, cheaper, and possibly more discriminating prescriptions
In this series there was no difference in the number or duration of hospital admissions between the two groups of doctors
The provision of 24 hour care for patients has become unacceptable to many general practitioners in the United Kingdom.1 The pressures of out of hours work of fatigue,2 concerns about personal safety,3 and a perception that many calls are inappropriate4 5 6 have been increased by the substantial rise in workload at night.7 As a result general practitioners delegate more out of hours care to other agencies, principally commercial deputising services. In 1964 deputising services were used by 9% of general practitioners.8 In 1993 they carried out over one third of night visits nationally and over two thirds of night visits in inner city areas.9 One outcome of recent negotiations with government is that personal care out of hours by general practitioners may further decrease.10
The advantages and disadvantages of deputising services have been described.11 12 13 14 15 We report what we believe to be the first controlled comparison of care provided by practice doctors and doctors from deputising services. This paper is concerned with process measures–namely, the response to the request for care, the time taken to visit at home, details of prescribing, and the proportion of patients admitted to hospital. Our accompanying paper examines outcome measures.16
Subjects and methods
The study took place between July 1994 and July 1995. Fourteen undergraduate teaching and postgraduate training practices were invited to participate, serving inner city and suburban populations in Manchester, Salford, Stockport, and Leicester. Each area had a different deputising service. For duty periods studied, participating practice doctors agreed either to provide out of hours care personally or to use a deputising service as determined by the randomisation process.
An “out of hours contact” was defined as any request for medical care between 7 pm and 7 am on weekdays, from 1 pm on Saturdays, and from 7 am for 24 hours on public holidays. A “night contact” was defined as any request for medical care between 10 pm and 7 am.
Duty periods were stratified to include a proportionate number of weekday evenings and nights, weekends, and bank holidays and then randomly allocated to care provided by either deputising services or practice doctors. Neither group of doctors was informed of which calls were being studied. Patients contacted their practice for care as usual throughout. Patients resident for less than two weeks were excluded, and out of hours contacts for a patient occurring within two weeks of the first contact were included as follow up calls. Patients who requested care during a study period were interviewed 24 to 120 hours later (two thirds between 24 and 72 hours). The average time to interview was equivalent in each arm of the study.
A structured questionnaire was used to collect sociodemographic data, time of request and time of visit if received, hospital admissions, satisfaction, health status, and details of drugs prescribed or dispensed, including the number contained in an out of hours formulary.17 The doctors studied were not aware that a formulary would be used as an indicator of prescribing. Satisfaction, health status, and subsequent use of health services are described in our accompanying paper.16 For children and patients unable to complete the interview, data were collected from a close relative. Interpreters were available when required.
Data collected from doctors' records included clinical information, treatment, and nature of the encounter. Twelve practices used the deputising service to pass calls to the doctor on duty, one used a local ambulance service, and one received calls direct. Details of calls were transcribed by practice doctors on to standard call sheets. This record of the time of the call and, for those patients visited, the time of arrival at the patient's home reported by the doctor was used in the analysis. Clinical problems and prescriptions were coded by using the read code system. Prescribing costs were drawn from the British National Formulary, Drug Tariff, or Mims current at the start of the study. The six interviewers underwent joint training and carried out quality standardisation procedures. All data were double coded, double entered, and verified.
Most variables showed evidence of correlation in outcome among patients seen by the same general practitioner or deputising doctor (intracluster correlation). As significance tests without adjusting for this variability between doctor and patient would be inappropriate, data were analysed by using a multilevel model.18 This was carried out with the MLn software package.19 “Adjusted percentages” indicate multilevel modelling with fitted values derived from a model with patients at the first level and doctors or practice at the second level, with adjustments for age, sex, and ethnic origin of the patient. Positively skewed variables were normalised by log transformation. 95% Confidence intervals were derived from the multilevel analysis. For proportions, confidence intervals are derived from the detransformed logistic regression estimates.
A total of 2152 calls were studied, of which 1046 (49%) were dealt with by 49 practice doctors and 1106 (51%) by 183 deputising doctors. When information about the status of the deputy was available 98 out of 160 (61%) were local principals, and they cared for 480 out of 850 (56%) of the patients seen by deputising doctors. The study periods from the sampling frame yielded one in six of the total annual calls for the practices, which were equally distributed between the two groups. Overall, 873 calls were received during weekdays and 1279 during weekends or public holidays. When the time of the call was recorded 570 out of 2021 calls (28%) were night calls. The attending doctor was called to certify death on 48 occasions and the researchers were asked by the practices not to contact 41 patients. Among the remaining eligible sample, 1466 interviews were carried out, a response rate of 71%.
Response to request–Table 1 shows that practice doctors were much more likely than deputising doctors to offer telephone advice. Practice doctors were equally likely to give telephone advice during the day and evening (19.4%; 95% confidence interval 14.5% to 25.6%) as at night (17.8%; 12.4% to 24.7%).
Time from receipt of call to home visit–For patients visited at home the median and mean times to arrival for practice doctors were 35 and 55.4 minutes and for deputising doctors 52 and 65.9 minutes. Normalisation of the positively skewed data by log transformation suggested a significantly shorter waiting time for practice doctors (P<0.0001). After adjustment for age the ratio of the geometric mean time to visit was 1.39 (95% confidence interval 1.19 to 1.64). Practice doctors were more likely to visit within one hour and within two hours (table 2). Delays reported by patients (overestimated by a mean of 15 minutes for both groups) confirmed this.
PrescribingOf 1274 patients interviewed, 756 (59%) said they had received a prescription (table 3). Practice doctors were less likely than deputising doctors to issue a prescription. All patients, including those with respiratory illness but not those with gastrointestinal illness, were more likely to receive an antibiotic from deputising doctors. Practice doctors prescribed fewer items and a higher proportion of generic items. Consequently their prescriptions were less expensive. Prescriptions issued by practice doctors and, in particular, drugs dispensed were more likely to be from the out of hours formulary17 than those issued by deputising doctors. There were no significant differences in the indices of prescribing among deputising doctors whether or not they were active local principals in general practice.
Hospital admissions–Of patients interviewed, 118 had been referred to hospital. There were no significant differences in the numbers of patients referred to hospital, the numbers subsequently admitted, or the durations of admission (table 4).
This study showed differences in most measures of the process of care provided by practice doctors and deputising doctors. Though practice doctors gave less telephone advice and visited more often than previously reported,6 20 21 22 23 they visited more quickly than deputising doctors, who are required to visit when requested.24 As patients who attend an out of hours centre are also seen more quickly than those visited by deputising doctors,25 patients whose condition necessitates an urgent visit might receive a faster response if deputising services could provide a wider range of responses to requests for out of hours care. Similarly, though global response time targets for deputising services are set by some health authorities,26 the identification and prioritisation of patients needing urgent attention may be more appropriate.
The higher overall volume and cost and lower proportion of drugs prescribed or dispensed from the formulary may indicate less discriminating prescribing and the use of promotional drug samples by deputising doctors. However, because similar prescribing was found in deputising doctors irrespective of whether or not they were local principals this prescribing behaviour may relate more to the nature of the work than to the characteristics of the doctors themselves. Increased prescribing has been reported in settings in which the doctor does not know the patient.27 This may be inherent in any system of care provision that is distinct or separate from the practice, such as a general practitioner cooperative.
In summary, differences between practice doctors and deputising services were found in the response to the request for care, time to visit, and treatment given but not in admissions to hospital. Methodological considerations and the impact of these differences on the outcome of care measured in terms of health status, patient satisfaction, and subsequent use of health services are presented in our accompanying paper.16
Funding: MRC Health Services Research Board. Service support for the participating practices was provided by Trent and North Western Regional Health Authorities.
Conflict of interest: None.