Family Practice Vol. 16, No. 6, 566-572
© Oxford University Press 1999
Do better quality consultations result in better health? Relationship between quality of consultations and health status of patients with non-acute abdominal complaints in general practice
Department of General Practice/Family Medicine, Utrecht University, Universiteitsweg 100, 3584 CG Utrecht, The Netherlands.
van Berkestijn LGM, Kastein MR, Lodder A, de Melker RA and Bartelink M-L. Do better quality consultations result in better health? Relationship between quality of consultations and health status of patients with non-acute abdominal complaints in general practice. Family Practice 1999; 16: 566–572.
Received 17 December 1998; Revised 21 May 1999; Accepted 25 June 1999.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
Background. In theory, a positive relationship is expected between the quality of a consultation and a patient's subsequent health status. However, such a relationship has not yet been firmly established in daily practice.
Objective. We aimed to study the relationship between the quality of the first consultation in a new episode of non-acute abdominal complaints and subsequent health status of patients in general practice.
Methods. Quality scores for 743 consultations were calculated on the basis of review criteria developed by expert panels. Functional health status was measured by the SIP (Sickness Impact Profile) at baseline, and at 1 and 6 months after the consultation. Multilevel regression analysis was used to examine the relationship between the quality of consultations and health status, and to identify factors of influence on this relationship.
Results. In the majority of these patients (97%) health status improved regardless of consultation quality. In patients with malignant disease, and chronic colitis, however, an association between consultation quality and subsequent health status was found: in those with a high consultation quality score (>66-percentile) the health status deteriorated in the first month but improved over the following 5 months; in those with a low consultation quality score (<33-percentile) it deteriorated continuously.
Conclusion. For the great majority of patients we found no relation between the quality of consultation and health status. However, for a very small subgroup of patients there is proof of benefit from better quality consultations.
Keywords. Abdominal complaints, consultation quality, general practice, health status..
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
It is important to assess the quality of GP performance in relation to patient outcomes.1,2 Little of the diagnostic process in general practice is concerned with acutely life-threatening conditions. Quite frequently the GP is confronted with patients presenting complaints of a non-acute, non-specific nature (e.g. abdominal complaints, fatigue, low back pain or headache).3 In these cases, the probability of organic disease is low, whereas missing such a diagnosis has serious consequences. This situation offers a good opportunity for assessing the quality of the diagnostic process in general practice.4,5
In this study we evaluated first consultations in a new episode of non-acute abdominal complaints. Although diagnostic management is often spread out over a series of successive consultations, rather than just one, much stress has deliberately been laid on the first consultation, based on the assumption that GP performance in the first consultation can influence the course of the complaints afterwards substantially.
Biometric parameters (e.g. blood test results) are not very appropriate as patient outcomes for this category of complaints, as they will hardly show abnormalities or clinically relevant changes: measuring improvement of health status is more suitable.6 The aim of this study was to assess how functional health status changes after the first consultation with a GP for a new episode of a non-acute abdominal complaint, and whether there is a relationship between the quality of this consultation and the subsequent functional health status of patients.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
Data collection on GP performance and patient health status
In a 6-month prospective study, data were collected on 743 consultations for non-acute, abdominal complaints carried out by 62 GPs.7,8 The latter were randomly selected from the total pool of GPs residing in or near cities with fewer than 100 000 inhabitants throughout The Netherlands, and constitute a representative sample. Criteria for inclusion of consultations were: first consultation for a new episode of abdominal complaints (any preceding episode lasting more than 3 months in the past), age of the patient between 18 and 75 years, and patient not referred immediately (the same day of the first consultation). Relevant data concerning history taking, physical examination, laboratory tests, imaging procedures and therapeutic management were collected. Activities were recorded from an analysis of case notes, combined with data from structured patient interviews, held within 24 hours of the consultation (t0) by a trained interviewer. This combination has previously proved to result in reliable information.9,10 The same interviewer also recorded a baseline health status measurement of the patients involved, and some patient and GP characteristics (Appendix 1). Second and third health status measurements were at 30 days (t1) and at 6 months (t2) after the first consultation by sending the patients a postal questionnaire to complete. Functional health status was measured by means of the Sickness Impact Profile (SIP), which has recently been validated.11–13 The SIP is a multidimensional general health status index that measures changes in physical, mental and social behaviour, ascribed by the patient to a current health disorder. The relative importance of each item of behavioural dysfunction in the index is expressed by a weighted score. The SIP score is calculated by summing the scores of the dysfunctional items answered positively. The SIP score ranges from 0 to 100. The higher the score, the worse the functional health status. The validity and reliability of the SIP in measuring the impact of an illness on performance of everyday activities are well established.14,15
Patients and diagnostic categories
The nature and content of the final diagnosis were independently assessed by two senior GPs of our department by examining all medical records 1 year after entry into the study. Subsequently, three diagnostic categories were constructed. A diagnosis was classified as ‘malignant/chronic’ when it was a malignant or chronic disorder (i.e. gastrointestinal malignancy, and Crohn's or ulcerative colitis). A diagnosis was classified as ‘proven organic’ when confirmed by laboratory test or imaging procedure, with the exception of the ones already classified in the category ‘malignant/chronic’. All the remaining diagnoses ware classified as ‘other’.
From the total of 743 patients, 37 (4.9%) were classified as ‘proven organic’. Of the 20 (2.6%) patients in the category ‘malignant/chronic’, six died within the follow-up period. One patient in the category ‘other’ died of myocardial infarction. Diagnoses encountered in the category ‘proven organic’ were (n = 37): duodenal ulcer (8), gastric ulcer (5), cholelithiasis (6), diverticulitis (2) and miscellaneous (16). Diagnoses encountered in the category ‘malignant/chronic’ were (n = 20): colorectal malignancies (7), pancreatic carcinoma (3), GIT metastases of unknown origin (3), gastric carcinoma (2), ulcerative colitis (4) and Crohn's disease (1).
Quality assessment of consultations
The quality criteria for the assessment of consultation quality in this study were constructed by means of the Delphi technique.16,17 The core of this technique consists of consultation of an expert panel by subsequent postal questionnaires. The entire Delphi process yields different lists of GP activities, appropriate for different patient situations. GP activities on these lists are prioritized by weight (score on a ten-point scale). Next, a Quality score was constructed, reflecting the quality of a consultation.18 The Quality score (the ratio of the sum of weights of appropriate activities carried to the sum of weights of all appropriate activities, multiplied by 100) represents the extent to which appropriate activities were carried out (the higher the score, the higher the quality rating). In order to classify consultations in classes of high, moderate and low quality, tertiles of the Quality scores were computed (high quality: Quality score >33-percentile; low quality: Quality score <33-percentile).
|
Analysis |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
In our study, the possible influence on the SIP scores of patient and GP characteristics was first analysed by bivariate analysis and all variables showing a significant bivariate relation were tested by multilevel regression analysis. Multilevel analysis is applied when variables stem from two or more hierarchical levels.19–22 Repeated measures in the same individuals belong to a lower level, computed means of subgroups sharing a common characteristic, to a higher level. In this study, data have a hierarchical structure in the sense that data from the natural level of measurement (the consultation/patient level) are aggregated at a higher level (the GP level). Also, data are disaggregated to a lower level: health status scores of patients, measured at different intervals after the first consultation, are repeated measures in the same individual. Aggregation and disaggregation of data to a hierarchically higher or lower level generates interdependency, since they are by then linked to each other by a common characteristic (in this study, the GP). Traditional regression analysis presupposes the independency of observations. Multilevel analysis allows for the dependency of measurements in these hierarchically structured data, and separately examines the effects of variables from different levels. Analysis of data of hierarchical levels by traditional regression analysis could produce spurious relations; separate examination by level prevents this type of error.
In order to assess the relationship between consultation quality and patient health status, we determined whether the general trend in change of SIP scores was affected by the Quality scores, while correcting for relevant patient and GP characteristics.
To illustrate relations between process and outcome for distinct subgroups of patients, mean SIP scores were plotted in diagrams.
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
Changes in health status
An overview of the observed SIP scores for different subgroups at the three observation times is shown in Table 1
. The observed SIP scores at t0, t1 and t2 are plotted in Figure 1. In the short term (between t0 and t1) the health status of the total population significantly improved, deteriorating again very slightly and not significantly during the second observation period (Fig. 1A). The changes in health status scores for patients in the diagnostic categories ‘proven organic’ and ‘other’ showed overall the same pattern as the total population, while the health status of patients in the diagnostic category ‘malignant/chronic’ deteriorated over both observation periods (Figure 1B).
|
|
The health status of patients with a ‘malignant/ chronic’ disorder and a high consultation Quality score (>66-percentile) deteriorated in the short term but improved again over the second period; the health status of those with a low consultation Quality score (<33-percentile) deteriorated continuously (Figure 1C
).
Factors of influence on health status
Patient characteristics with significant influence on health status in the first consultation were: health perception of the patient at t0, perceived inconvenience of the complaints at t0, perceived seriousness of the complaints at t0, long duration of the complaints at t0, presence of comorbidity, presence of psychosocial problems and being a frequent attender. None of the GP characteristics investigated showed significant influence on health status.
Consultation quality and health status
No significant effect of the Quality score was found to explain the initial improvement in health status in patients in the diagnostic categories ‘proven organic’ and ‘other’. Only in the very small subgroup of patients with malignant disease, and Crohn's or ulcerative colitis, was a significant (P < 0.001) relationship between health status and consultation quality found: the more these patients were subjected to appropriate activities, the better they did, while the health status of those exposed to fewer appropriate activities deteriorated, as already mentioned, steadily. This relationship between consultation quality and health status remained significant (P < 0.001) after correction for the effects of significant patient characteristics.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
Main results
The substantial and enduring improvement of the health status of the great majority of the study population appeared to be independent of the level of appropriate medical activities in the first consultation. This finding is in agreement with that of another primary-care study, in which no relationship was found between the adequacy of history taking, physical examination, use of diagnostic tests or prescription of drugs in consultations for a wide variety of common non-acute disorders and the resolution of the patient's symptoms.23
Only in a very small subgroup of patients was there a relationship between consultation quality and health status: the more patients with a malignant disease and colitis were subjected to appropriate activities, the better their health status in the long term. Though the numbers of patients involved were very low, the correlation appeared to be significant (P < 0.001).
Lack of effect of appropriate medical activities in the majority of patients
As in another Dutch study on non-acute abdominal complaints, there was little ‘proven organic’ disease in our study population (4%).24 This could in itself explain why we found no effect of the level of appropriate activities on health status in the total population. In general practice, many disorders are self-limiting and/or not amenable to treatment. Most of the patients that visit a GP tend to improve rapidly no matter what the doctor does. This assumption is consistent with other studies on non-specific complaints in general practice that have found no relationship between the doctor's actions and the resolution of complaints.23,25 Patients go to see their GP with an annoying symptom at the moment it reaches its highest peak, so the GP is confronted with a selection of people with complaints that are extreme compared with the general population. In these patients a second health status measurement after a short time therefore shows health status scores improving towards the mean of the ‘healthy’ population.
Effect of appropriate medical activities in patients with a ‘malignant/chronic’ disorder
For patients with a ‘malignant/chronic’ disease the situation is different. The Delphi panel experts considered patient background data crucial to the assessment of the appropriateness of diagnostic activities by GPs. For instance: more questions needed asking and more diagnostic activities were required in older (>40 years) patients, while in younger patients some questions were unnecessary and several investigations were judged to be superfluous. One of the striking results from this study was that specific enquiries about alarming symptoms were made in fewer than half the consultations with older patients. From a preliminary analysis of the same study population it is known that the initial complaint of the majority of patients was restricted to abdominal pain, bloatedness and/or flatulence, and that alarming symptoms were seldom reported spontaneously.2617 Differences in quality of consultations were mainly due to the extensiveness of history-taking.11 The positive effect of appropriate activities (= more extensive history-taking) on health status could be due to timely detection of malignancy and consequently earlier diagnosis and treatment.27–30 This could be an explanation for the relationship between consultation quality and health status in patients with a malignant/chronic disorder. This assumption could be convincingly confirmed by studying the medical records of the 20 patients concerned.
Limitations of the study
The findings of this study only apply to first consultations in a new episode of non-acute abdominal complaints in general practice. The effect of a series of consultations was not studied. The results are not applicable to acute abdominal disorders in which there may be more immediate vital risk and/or totally different probabilities of disease.
In this study we considered only the purely medical aspects of GP performance. The effects of interpersonal skills of the GP were not investigated, because this would make a too complicated study design. In common disorders, particularly, discussion and agreement between the patient and physician about the nature of the patient's problem has been proven to have a positive influence on the outcome of care.31–33 Reaching such agreement calls for use of interpersonal skills by the GP. In quality assessment of the management of common disorders it could therefore be rewarding to assess GPs' interpersonal skills.
|
Conclusions |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
Better quality consultations do only result in better health status in a very small subgroup of patients: those with malignant disease and chronic colitis. The great majority of patients with other diagnoses improve substantially, but independently of consultation quality.
Especially in cases of non-acute abdominal complaints with a low prior probability of organic disease, GPs should pay special attention to identifying patients at increased risk of gastrointestinal malignancy. The importance of such a prognostic approach has been stressed by several authors.34–36 Moreover, information from history-taking has a higher predictive value than the results of laboratory tests,37–40 which underlines the importance of good history-taking.
It has always been the mission of primary care to optimize early diagnosis in organic diseases. One of the most difficult things in good clinical practice is to find, among all the patients consulting us, the ones to be picked out for further investigation.
The fact that better history-taking only results in better health in a very small subgroup of patients does not relieve the GP from keeping alert on alarming symptoms. GPs should take the initiative and question patients over 40 years of age on alarming symptoms and risk indicators. Failing to ask such questions results in incomplete data and misjudgement of the probability of malignant disease. Little time is required to ask some extra questions during history-taking, while the benefit for the patient can be substantial.
|
Appendix:Overview of variables used in this study |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
|
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsAnalysisResultsDiscussionConclusionsAppendix:Overview of variables...References |
|---|
1 Huygen FJ, Mokkink HG, Smits AJ, Son JA van, Meyboom WA, Eyk JT van. Relationship between the working styles of general practitioners and the health status of their patients. Br J Gen Pract 1992; 42: 141–144.[Web of Science][Medline]
2 Headache Study Group of the University of Western Ontario. Predictors of outcome in headache patients presenting to family physicians. Headache 1986; 26: 285–294.[Web of Science][Medline]
3 Lamberts H, Brouwer HJ, Mohrs J. Reason for Encounter- episode- and Process-oriented Standard Output from the Transition Project. Amsterdam: Department of General Practice, 1990.
4
Kuyvenhoven MM, Jacobs HM, Touw-Otten FWMM, Es JC van. Written simulations of patient-doctor encounters. Research instrument for registration of the performance of general practitioners. Fam Pract 1984; 1: 14–29.
5 Pieters HM. Simulated patients in assessing consultation skills of trainees in general practice vocational training: a validity study. Med Educ 1994; 28: 226–233.[Web of Science][Medline]
6 Donabedian A. Explorations in Quality Assessment and Monitoring. Volume I. The Definition of Quality and Approaches to its Assessment. Ann Arbor, Michigan: Health Administration Press, 1980.
7
Berkestijn LGM van, Kastein MR, Lodder A, Melker RA de, Bartelink ML. How well are patients treated in general practice. Quality of consultations for non-acute abdominal complaints. Int J Qual Health Care 1998; 10: 221–233.
8 Jacobs HM, Luttik A, Touw-Otten FWMM, Kastein MR, Melker RA de. Measuring impact of sickness in patients with nonspecific abdominal complaints in a Dutch family practice setting. Med Care 1992; 30: 244–251.[Web of Science][Medline]
9 Gerbert B, Hargreaves WA. Measuring physician behavior. Med Care 1986; 24: 838–847.[Web of Science][Medline]
10 Romm FJ, Putnam SM. The validity of the medical record. Med Care 1981; 19: 310–315.[Web of Science][Medline]
11 Jacobs HM, Luttik A, Touw-Otten FWMM, Melker RA de. De 'Sickness Impact Profile’. Resultaten van een valideringsonderzoek van de Nederlandse versie. (The ‘Sickness Impact Profile’. Results of a validity study of the Dutch version of the SIP) (in Dutch with summary in English). Ned Tijdschr Geneesk 1990; 134: 1950–1954.
12 Jacobs HM. Health status measurement in family medicine research. The Sickness Impact Profile and its application in a follow-up study in patients with non-specific abdominal complaints (dissertation). Utrecht: Utrecht University, 1993.
13 Bergner M, Bobbitt RA, Pollard WE, Martin DP, Gilson BS. The Sickness Impact Profile: validation of a health status measure. Med Care 1976; 14: 57–67.[Web of Science][Medline]
14 Pollard WE, Bobbitt RA, Bergner M, Gilson BS. The Sickness Impact Profile: reliability of a health status measure. Med Care 1976; 14: 146–155.[Web of Science][Medline]
15 Deyo RA, Diehr P, Patrick DL. Reproducibility and responsiveness of health status measures: statistics and strategies for evaluation. Cont Clin Trials 1991; 12: 42S–58S.
16 Kastein MR, Jacobs HM, Hell RJ van der, Luttik A, Touw-Otten FWMM. Delphi, the issue of reliability. A qualitative Delphi study in primary health care in the Netherlands. Technol Forecast Soc Change 1993; 44: 315–323.
17 Ashton CM, Kuykendall DH, Johnson ML et al. A method of developing and weighting explicit process of care criteria for quality assessment. Med Care 1994; 32: 755–770.[Web of Science][Medline]
18 Kastein MR. Developing criteria for the evaluation of performance in family medicine using the Delphi technique (dissertation). Utrecht: Utrecht University, 1994.
19
Korff M von, Koepsell Th, Curry S, Diehr P. Multi-level analysis in epidemiologic research on health behaviors and outcomes. Am J Epidemiol 1992; 135: 1077–1082.
20
Diez-Roux AV. Bringing context back into epidemiology: variables and fallacies in multilevel analysis. Am J Publ Health 1998; 88: 216–222.
21 Hox JJ, Kreft IGG. Multilevel analysis methods. Sociolog Methods Res 1994; 22: 283–299.
22 Witte JS, Greenland S, Kim LL. Software for hierarchical modeling of epidemiological data. Epidemiology 1998; 9: 563–565.[Web of Science][Medline]
23 Bass MJ, Buck C, Turner L, Dickie G, Pratt G, Robinson HC. The physician's actions and the outcome of illness in family practice. J Fam Pract 1986; 23: 43–47.[Web of Science][Medline]
24 Muris JWM, Starmans R. Non-acute abdominal complaints. Diagnostic studies in general practice and outpatient clinic (dissertation). Maastricht: University of Limburg, 1993: 106.
25 Thomas KB. The consultation and the therapeutic illusion. Br Med J 1978; 121: 1327–1328.
26 Jacobs HM, Luttik A, Melker RA de, Hell RJ van der, Kastein MR, Touw-Otten FWMM. Patiënten met niet-specifieke buikklachten en huisartsgeneeskundig handelen (Primary care in patients with nonspecific abdominal complaints) (in Dutch, with summary in English). Tijdschr Soc Geneesk 1993; 71: 79–86.
27 Hampton JR, Harrison MGJ, Mitchell JRA, Prichard JS, Seymour C. Relative contributions of history taking, physical examination and laboratory investigation to diagnosis and management of medical outpatients. Br Med J 1975; II: 486–489.
28 Sandler G. The importance of the history in the medical clinic and the cost of unnecessary tests. Am Heart J 1980; 100: 928–931.[Web of Science][Medline]
29 Sackett DL. Clinical Epidemiology. New York: Little Brown, 1985: 17.
30
Pryor DB, Shaw L, McCants CB et al. Value of the history and physical in identifying patients at increased risk for coronary artery disease. Ann Int Med 1993; 118: 81.
31 Bass MJ, Buck C, Turner L, Dickie G, Pratt G, Robinson HC. The physician's actions and the outcome of illness in family practice. J Fam Pract 1986; 23: 43–47.
32 Starfield B, Wray C, Hess K, Gross R, Birk PS, D'Lugoff BC. The influence of patient-practitioner agreement on outcome of care. Am J Public Health 1981; 71: 117–131.
33 Stewart MA, McWhinney IR, Buck CW. The doctor-patient relationship and its effect upon outcome. J R Coll Gen Pract 1979; 29: 77–82.[Medline]
34 Numans ME, Graaf Y van der, Wit NJ de, Touw-Otten FWMM, Melker RA de. How much ulcer is ulcer-like? Diagnostic determinants of peptic ulcer in open access gastroscopy. Fam Pract 1994; 11: 208–216.
35 Colin Jones DG. Management of dyspepsia: Report of a working party. Lancet 1988; i: 576–579.
36
Fijten GH, Starmans R, Muris JWM, Schouten HJA, Blijham GH, Knottnerus JA. Predictive value of signs and symptoms for colorectal cancer in patients with rectal bleeding in general practice. Fam Pract 1995; 12: 279–286.
37 Hampton JR, Harrison MGJ, Mitchell JRA, Prichard JS, Seymour C. Relative contributions of history taking, physical examination and laboratory investigation to diagnosis and management of medical outpatients. Br Med J 1975; II: 486–489.
38 Sandler G. The importance of the history in the medical clinic and the cost of unnecessary tests. Am Heart J 1980; 100: 928–931.
39 Sackett DL. Clinical Epidemiology. New York: Little Brown, 1985: 17.
40 Pryor DB, Shaw L, McCants CB et al. Value of the history and physical in identifying patients at increased risk for coronary artery disease. Ann Int Med 1993; 118: 81.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||