Reasons for encounters, investigations, referrals, diagnoses and treatments in general practice in Sweden—a multicentre pilot study using electronic patient records

Abstract

Objective: To investigate reasons for encounters, investigations, referrals, diagnoses and treatments in everyday general practice, using electronic patient records (EPR), and possible related differences concerning gender, socio-economic status (SES) and practice location. Method: Four Swedish primary care centres using EPR participated. Distributions of symptoms, investigations, diagnoses and prescribed drugs were registered. Results: In 1055 encounters, the mean patient age was 53; 59% were women. The most common reasons for the encounter were musculoskeletal (21.5%) and respiratory (15.2%) symptoms. A total of 1534 diagnoses were coded, on average 1.5 per encounter. The predominant diagnostic groups, i.e. ICD-10 chapters, were musculoskeletal (17.2%) and respiratory (12.4%). The most common specific diagnoses were essential hypertension (8.1%) and acute upper respiratory infections (3.7%). A total of 1687 prescriptions were issued, on average 1.6 per encounter. The most frequent pharmaceutical groups were nervous (17.7%), respiratory system (16.2%), and cardiovascular (15.7%). The most frequent drugs were phenoxymethyl penicillin (3.7%), diclofenac (2.9%) and acetylsalicylic acid (2.5%). An average of 1.3 laboratory tests was performed per encounter. In 7.5% of encounters, radiology referrals were made; in 12.3% referrals were made to other specialists/therapists, while sick-list certificates were written in 11.7%. There were significant differences concerning symptoms, diagnoses and investigations between female and male patients, urban and rural practices and SES.

Conclusion: The musculoskeletal, respiratory and circulatory systems predominated, both as reasons for the encounter and in the diagnoses, but with significant differences concerning gender, SES and practice location.

Key words::

• Primary health care
• general practice
• electronic patient record
• reasons for encounter
• symptoms
• health problem
• prescribing

Introduction

In Sweden, general practitioners (GPs) conduct about 13 million consultations per year (1), and the use of electronic patient records (EPR) is now almost universal (2,3). This makes it easier to retrieve clinical information, but difficulties and limitations are involved (4). The accuracy (correctness and completeness) of clinical data in such records can be considered to be high (5,6) while socio-economic data are generally documented more poorly. Use of a primary care version of the ICD-10 diagnostic system is generally accepted among GPs in Sweden (7), and the widespread use of computers in prescribing, ordering laboratory procedures and in handling referrals increases the possibility of retrieving clinical data from records in everyday clinical practice (8). In spite of this, statistics from primary health care (PHC), especially focusing on symptoms, have been sparse (9) and there is a need for more information about health problems and management in general practice (10). We wanted to explore if it is possible to describe everyday general practice from a structured collection of data from EPRs for quality improvement and research.

More extensive data on the ‘episode of care,’ using the International Classification for Primary Care (ICPC) for collecting initial data in the diagnostic process, has been presented by Yamada et al (10). Furthermore, a comparison of the content in family practice in different countries (Netherlands, Japan, US and Poland) was published in 2002 (11); data collected from EPR covering reasons for encounters, diagnoses and interventions in primary care in the different countries were presented. Both similarities and differences were found. The differences primarily found concerned the utilization of care per episode, including diagnoses and therapeutic interventions (11). These data were based on population and annual encounter data. Data on utilization, reasons for encounter per episode of care and prescriptions were studied. The authors also recommended studies based on standardized documentation of episodes of care using the ICPC mapped to the ICD-10.

Our hypothesis was that comparable data, suitable for research purposes, could be collected from structurally and uniformly collected EPR data at primary health care centres (PHCC) for studying possible differences in management with respect to gender, socio-economic status (SES), and practice location.

The aim of this study was to investigate reasons for encounters and to study, in relation to the presented symptoms and based on consecutive data in EPR, the diagnoses, investigations, prescriptions and referrals in everyday general practice, as well as to investigate possible differences related to gender, SES and practice location.

Methods

Design

Cross-sectional, evaluative and comparative study of data withdrawal during different time intervals at four PHCC in Sweden. The Ethics Committee at the University of Gothenburg approved the study and participants gave informed consent.

Health centres

12 GPs at four PHCC in the south-western region of Sweden agreed to participate (12). The centres were strategically selected to obtain a mixture of private and public funding, urban and rural location and different EPR systems, to be able to test our hypothesis concerning comparability for research purposes. Two of the four PHCC were privately financed and two were publicly financed. Two PHCC were rural and two were urban. The GPs collected data regularly, during either one or two days per week or entire weeks, during parts of the period from January to September 2001 (the summer holiday period 15 June to 15 August was excluded). The data collection periods were determined in advance for each PHCC, based on organizational and staffing factors.

Patients

All patients visiting the GP on study days were asked to participate and received written information about the study. Patients who agreed to participate gave their written permission; those who declined were not registered. The PHCC were contacted regularly during the study period to ensure uniform registration of factors such as diagnostic codes.

Structured data entry and variables

The encounter and registration structure of all encounters was based on the subjective-objective-assessment-plan (SOAP) structure (13).

Subjective. In the Subjective part, structured data were exclusively collected, prior to the encounter, in a form given to the patient that included information on: (1) reasons for the encounter; (2) first visit or re-visit for the same ailment; (3) social variables— such as marital status, number of children, profession and smoking (0–10 cigarettes per day or > 10 cigarettes per day); and (4) alcohol (not at all, moderately or frequently). Secretaries entered this information into the EPR. The reasons for encounter, presented as symptoms by the patients in the questionnaires, were, independently and temporally separated from the encounters, coded according to the ICPC-2 classification (14) by one of the authors (JM) without knowledge of the ICD codes of the respective encounters.

Objective. In the objective part, variables normally included in a physical examination by a GP were registered, with an additional dichotomized code (normal or pathological).

Assessment. In the Assessment part, diagnostic codes were registered by the GPs using the Swedish PHC version of the ICD-10 (7). The three most important codes for the encounter in question were registered for each visit.

Plan. In the plan part, the number of laboratory tests, radiological referrals, referrals to other specialists and therapists, prescriptions (coded according to the Anatomic Therapeutic Chemical (ATC) classification (15)) and sick-list certificates were automatically registered.

Data retrieval and database

Three of the most common EPR systems in Sweden (Profdoc^®, BMS^® and Biosis^®) were used at the centres. The data were retrieved using online searches, based on standard question language (SQL), via the statistical modules in the EPR systems.

Assessment of data accuracy

Before data collection began, all participating GPs took part in a joint quality assurance process concerning diagnostic coding. To ensure uniform registration, four meetings were held and continuous contact between one of the authors (JM) and the participants was maintained during the study period.

The centres were contacted regularly during the study period so that the procedure could be continuously monitored and evaluated. We compared age and gender distributions in the study encounters with those in all encounters in all PHCC in order to assess representativeness

All registrations at two centres, and random samples consisting of one in ten registrations at the other two centres, were evaluated by one of the authors (JM) to ensure that the information in the retrieved data was correct and complete.

Variables created for the present study

A SES variable was defined for this study, based on occupation, which was defined either as lower-level: (i.e. blue-collar workers and lower-level white-collar workers) or higher-level (i.e. higher-level white-collar workers and businesspeople). A PHCC located in a town or city was denoted as urban, while a PHCC located in a village was denoted as rural.

Statistical analysis

Data were processed using SPSS^® for statistical analysis. Frequencies of symptom codes, diagnostic codes and ATC-codes were presented according to the total number of codes (n), respectively. The chi-square test was used for comparison between different categories. The significance level was set at 0.05.

Results

Subjects and encounters

We recorded 1055 encounters at four PHCC, of which 59% involved female patients; 54% of the encounters were re-visits for the same ailment. The mean age of the patients was 53 years. The distributions of age and gender in the study population did not differ significantly from those in the corresponding patient populations at the PHCC.

Subjective data

A total of 1346 ICPC-2 codes were registered at the PHCC, yielding an average of 1.3 codes per encounter. These 1346 ICPC-2 codes indicated that the most frequent reasons for an encounter were musculoskeletal symptoms (21.5%), respiratory symptoms (15.2%) and circulatory symptoms (13.3%) (Table I). Women were predominant in the eight most common groups, ranging from 53.6 to 71.6%, with ‘general’ and ‘unspecified symptoms’ as the most common main group of symptom codes among women.

Table I. Subjective part: distribution of symptom codes (n = 1346) in 1055 encounters, with a frequency of > 1.0% across ICPC-2 chapters.

ICPC-2 chapter	% of symptom codes
Musculoskeletal (locomotion)	21.5
Respiratory	15.2
Circulatory	13.3
Psychological	8.0
Endocrine, metabolic and nutritional	7.1
Digestive	7.0
Skin	6.7
General and unspecified symptoms	6.0
Ear (hearing)	3.0
Standard process components of ICPC ^a	2.9
Neurological	2.5
Urological	2.4
Others	4.4
Total	100

^aIncluding medical examination, partial medical examination, clarification, medication, repair/fixation, therapeutic counselling, results tests/procedures and administrative procedure.

In 54% of cases the patients’ and physicians’ health assessments concurred, while the difference was greater than four percentage points in 3%; 76% in this latter group were women with a mean age of 40.

Objective data

Sixty-two percent of the encounters (n=1055) resulted in at least one notation of a pathological finding. In 58% of the visits, a comment on general health was noted in the EPR (Table II). Heart auscultation was performed in 23% of the encounters.

Table II. Physical examinations (n = 1816) performed in 1055 encounters, in descending frequency, and the proportion of abnormal findings.

Examination	Frequency of registered physical examinations in relation to encounters (%)	Proportion of abnormal findings in relation to registered physical examinations (%)
General condition	58	18
Local examination	25	88
Heart	23	12
Mouth and throat	15	58
Lymph nodes	11	14
Abdomen	9	32
Ears	9	46
Thyroid gland	4	11
Lungs	2	62
Rectal palpation	2	62
Neurology	2	32
Eyes	2	65
Sinuses	2	50
Breasts	<1	0

Assessment and diagnosis

A total of 1534 diagnoses were coded, an average of 1.5 per encounter. The predominant diagnostic groups (i.e. ICD-10 chapters) were diseases of the musculoskeletal system (17.2%), respiratory system (12.4%), and circulatory system (11.6%) (Table III). The most common specific diagnoses were essential hypertension (8.1%) and acute upper respiratory infections (3.7%).

Table III. Distribution of diagnostic codes (n = 1534) in 1055 encounters, with a frequency of >1.0% across ICD-10 chapters.

ICD-10 chapter and diagnosis	Percentage of diagnostic codes
Diseases of the musculoskeletal system and connective tissue	17.2
Diseases of the respiratory system	12.4
Diseases of the circulatory system	11.6
Factors influencing health status and contact with health services e.g. observation for suspected diseases or for administrative purposes	9.8
Mental and behavioural disorders	9.3
Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified	8.3
Endocrine, nutritional and metabolic diseases	5.7
Diseases of the genito-urinary system	4.6
Diseases of the skin and subcutaneous tissue	4.4
Injury, poisoning and certain other consequences of external causes	4.0
Diseases of the ear and mastoid process	3.5
Diseases of the digestive system	3.3
Certain infectious and parasitic diseases	3.1
Other	2.7
Total	100

Plan: laboratory tests, referrals and treatment

An average of 1.3 laboratory tests was performed per encounter. In 5.8% of the encounters only one laboratory test was performed, while more than 10 tests were performed in 23.4% of the encounters.

Referral to radiology occurred in about 7.5% of the visits, and in 12.3% a referral was sent to another specialist or therapist. In about 11.7% of the visits, the physician wrote a sick-list certificate.

A total of 1687 ATC-coded prescriptions were issued. The prescriptions were issued in 57% of the encounters, resulting in 2.8 prescriptions per encounter during which a prescription was issued. The most frequent main ATC groups were nervous system (17.7%), respiratory system (16.2%), and cardiovascular system (15.7%) (Table IV). The most frequent individual drugs were phenoxymethyl penicillin (3.7%), diclofenac (2.9%) and acetylsalicylic acid (2.6%).

Table IV. Distribution of prescriptions (n = 1687) in 1055 encounters, in descending frequencies across ATC main groups with a frequency of >1.0%

Main ATC group	Percentage of ATC codes in main group
Nervous system	17.7
Respiratory system	16.2
Circulatory system	15.7
Alimentary tract and metabolism	11.7
General anti-infectives for systemic use	10.1
Musculo-skeletal system	9.8
Dermatologicals	5.2
Blood and blood forming organs	4.5
Systemic hormonal preparations, excluding sex hormones	3.7
Sensory organs	2.6
Genito-urinary system and sex hormones	2.2
Other	0.6
Total	100

Differences in gender, SES and PHCC location

The most common symptoms, diagnoses, prescriptions, examinations and referrals varied in relation to patients´ gender and SES (as defined above), as well as to the urban or rural location of the PHCC. The distribution of symptoms in men differed significantly from that in women (Table V). Men more often had musculoskeletal symptoms and were given prescriptions for anti-infectious drugs, while women more often had circulatory symptoms.

Table V. Differences in the six most common codes for first reported symptoms and diagnoses for each encounter, all prescriptions, radiological examinations and referrals in relation to gender, urban/rural PHCC location and socio-economic status (SES) (based on profession).

		Gender			Urban/rural PHC		SES
		n (%/χ)	♂ %	♀ %	Urban %	Rural %	High %	Low %
Symptom	Circulatory	213 (20)	22	33	28	31	21	38
codes	Musculoskeletal	173 (16)	27	22	20	31	31	22
(ICPC-2)	Respiratory	131 (12)	20	17	19	14	21	9
n = 1055	Psychological	83 (8)	10	12	14	6	10	12
	Digestive	75 (7)	12	9	10	11	11	11
	Endocrine	59 (6)	9	7	9	7	6	8
		734 (70)	100	100	100	100	100	100
			P = 0.02		P = 0.003		n.s.
Diagnostic	Musculoskeletal	262 (25)	22	30	26	29	21	33
codes	Respiratory	191 (18)	21	19	16	27	27	22
(ICD-10)	Circulatory	186 (18)	21	18	20	18	17	11
n = 1055	Mental/behavioral	112 (11)	11	12	12	10	6	13
	Factors infl. health	111 (11)	12	11	15	5	20	11
	Symptoms, signs	109 (10)	13	10	11	11	9	10
		971 (92)	100	100	100	100	100	100
			n.s		P = 0.001		P = 0.02
Prescription	Nervous	299 (28)	21	22	22	22	15	21
codes	Respiratory	273 (26)	20	20	22	17	21	20
(ATC)	Cardiovascular	265 (25)	19	20	18	21	15	8
n = 1687	Alimentary tract	197 (19)	16	14	13	17	7	12
	Anti-infective	171 (16)	15	10	11	14	30	18
	Musculoskeletal	165 (16)	9	14	14	9	12	21
		1370 (81)	100	100	100	100	100	100
			P = 0.03		P = 0.001		n.s.
X-ray		79	2	2	4	0	5	5
			n.s.		p < 0.0001		n.s.
Referrals to other specialists		130	4	3	7	0	7	7
			n.s.		P < 0.0001		n.s.

χ,consultation.

In rural PHCC, patients most often presented with musculoskeletal symptoms and were given respiratory diagnoses, while respiratory symptoms, ‘factors influencing health status’ and radiology and other referrals were significantly more common in urban PHCC.

Discussion

Main results

In this study, we investigated the reasons for encounters as well as the diagnoses, investigations and referrals in these everyday general practices, in relation to presented symptoms. The musculoskeletal, respiratory and circulatory systems predominated, both as reasons for the encounter and in the subsequent diagnoses. The most common pharmacological treatment was for nervous and respiratory diseases and the most common pharmaceutical drug was phenoxymethylpenicillin. Significant differences were found concerning the six most common symptoms, diagnoses, prescriptions, examinations and referrals in relation to PHCC location, gender and SES.

Strengths and limitations

An advantage of our study was the strategic selection of the four studied PHCC, resulting in a mixture of privately/publicly financed and rural/urban locations, which increased representativeness. Indeed, age and SES were fairly similar between the participating patients and the population in the PHCC catchment areas.

Since this study has been performed (2001), health reforms and reimbursement systems have changed, technology has developed and similar studies have started in Canada (16,17) and Denmark (18). Similar types of registrations have been made in the Netherlands, Belgium and Germany (11,19,20).

Registration in the EPR was structured in a standardized manner by all participating GPs to reduce individual variation. There are differences in coding between physicians, but the differences in terms of major ICD-10 chapters are probably minor. While difficult to assess, the accuracy of the collected information was probably high as the participants and researchers jointly and repeatedly confirmed the registration procedure and record structure.

Data collection was performed during different periods and not over a whole year, depending on conditions at the PHC centres, which makes this study less representative. However, the data was extracted during the first two thirds of the year, in a consecutive way, to minimize the seasonal variations with flues and viral infections in the late autumn, and to reduce influences of organizational character and tourist migrations during the vacation period.

All encounters were registered individually, even though 54% were re-visits. Studying each individual patient independently might have been better. Another limitation was that the patients who declined participation were not registered. The relatively small sample size was another major limitation of the study; larger studies with continuous collection of data from PHCC are required in the future.

There was a possibility of certain definitions and measurement bias as the individual doctor was aware of the symptoms reported by the patient. The ICPC coding was, therefore, made independently by one of the authors without knowledge of the ICD codes.

Gender issues

The proportion of women in our study (59%) was quite similar to that in other studies, reporting 54–60% females (1,22). The symptom codes ‘general’ and ‘unspecified symptoms’ were common, emphasizing how difficult and resource-demanding it is for PHCC to investigate, plan for and pharmaceutically treat this large group of patients.

Comparisons between male and female consultation rates in different countries have previously been published by Fleming and Pavlic, finding striking similarities in male-female consultation rates concerning most ICD-10-chapter diagnoses (23).

Reasons for encounters

The main reasons for encounter in our study were symptoms in the musculoskeletal, respiratory and circulatory ICPC chapters (21.5–13.3%), which can be compared to another study in which the corresponding figures were 16.3–7.2% for the same chapters (24). When it comes to reasons for the encounters, our results are noticeably similar to those in an earlier publication on international comparisons between four countries. The conclusion in that study, confirmed by our study, indicates that it is the reasons for the encounters rather than the diagnoses that are most similar between countries (11).

ICDC-10 codes

The distributions of ICD-10 codes were in agreement with those obtained in other studies (4,5). The number of diagnostic codes per encounter (1.5) seems to be fairly high (despite the permitted maximum of three codes), but is in accordance with figures from other studies in which this number varied between 1.1 and 1.6 (5, 24). One Swedish study of particular interest (25) reported much lower rates of diseases of the circulatory system (7.7%) than in our study, while the rates of injuries and ophthalmological diseases were markedly higher in that study. When it comes to the frequency of single codes, our results were similar in some respects to results from similar studies (4,25,26). The most common single diagnostic code, essential hypertension, was, however, even more predominant in our study (8.1%), while accounting for only 2.9 to 6.4% in the studies mentioned above.

Examinations

The proportion of encounters entailing laboratory tests in our study (31%) can be compared to the corresponding figures from Iceland, ranging from 10.1% (rural) to 14.1% (urban) (27), and from Australia (12.8%) (24). In these studies, the proportion of encounters in which radiological examinations were ordered was 7.5%, and thus markedly higher than 2.5% (rural) to 1.5% (urban) (19) and 5.7% (25), reported in other studies. These studies also had lower rates of referrals to other therapists/specialists than in our study (12.3%), i.e. 1.5% (rural), 2.4% (urban) and 8.6%, respectively.

Prescriptions

Prescribing medication was generally frequent (57%), but comparable to Iceland - in which 50.7% (rural) and 47.9% (urban) of encounters resulted in a prescription (27) –and Australia, for which the corresponding rate was 60.3% (24). Prescriptions in the nervous and respiratory system main ATC groups were more frequent, while alimentary tract and circulatory system medications were somewhat less frequently prescribed than the rates reflected in national sales statistics (28). This may be as the latter includes all outpatient health care, including hospital-based outpatient care. Compared to figures on prescriptions for antibiotics from Australia (19.8%) (24), the rates we found were lower (11.9%).

Sickness certification

Sick-list certificates were more frequent than in other Swedish studies—11.7% compared to 9%—that might be due to regional or seasonal variation (29).

Implications for research and practice

This pilot study shows that there were significant differences between female and male patients, urban and rural locations and socioeconomic groups concerning the most common symptoms, diagnoses, prescriptions and investigations at these PHCC. This confirms both the importance of consultation skills and of the GP's ability to adapt his or her consultation mode to the individual patient (30). Furthermore, the importance of adapting reimbursement systems according to practice location and socioeconomic circumstances is also evident from our findings. However, the results should be confirmed both by larger national studies and by comparative studies including other countries.

Conclusion

The musculoskeletal, respiratory and circulatory systems are predominant in Swedish general practice, both as reasons for encounters and in the subsequent investigations, referrals, prescriptions and diagnoses. There appear to be significant differences between female and male patients, urban and rural locations and socioeconomic groups concerning symptoms, diagnoses, prescriptions and investigations.

Acknowledgements

This study was supported by grants from the National Board of Health and Welfare, and the Västra Götaland and Halland County Councils. The authors should like to thank Patrik Sjöberg from Tieto Enator and the staff at the following PHCC: Husläkarna i Kungsbacka, Kungsportsläkarna, Ytterby, Åsa.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.