Challenges and uncertainties with test result interpretation can lead to diagnostic errors. Primary care doctors are at a higher risk than specialists of making these errors, due to the range in complexity and severity of conditions that they encounter.
This study aimed to investigate the challenges that primary care doctors face with test result interpretation, and to identify potential countermeasures to address these.
A survey was sent out to 7800 primary care doctors in South Africa. Questionnaire themes included doctors’ uncertainty with interpreting test results, mechanisms used to overcome this uncertainty, challenges with appropriate result interpretation, and perceived solutions for interpreting results.
Of the 552 responses received, the prevalence of challenges with result interpretation was estimated in an average of 17% of diagnostic encounters. The most commonly-reported challenges were not receiving test results in a timely manner (51% of respondents) and previous results not being easily available (37%). When faced with diagnostic uncertainty, 84% of respondents would either follow-up and reassess the patient or discuss the case with a specialist, and 67% would contact a laboratory professional. The most useful test utilisation enablers were found to be: interpretive comments (78% of respondents), published guidelines (74%), and a dedicated laboratory phone line (72%).
Primary care doctors acknowledge uncertainty with test result interpretation. Potential countermeasures include the addition of patient-specific interpretive comments, the availability of guidelines or algorithms, and a dedicated laboratory phone line. The benefit of enhanced test result interpretation would reduce diagnostic error rates.
Laboratory services play an integral role in the healthcare system - from primary- through to tertiary-level care – since diagnostic tests can either confirm or exclude a tentative diagnosis, or screen for potential diseases.
Primary care doctors are usually the point of entry into a healthcare system, and are consequently exposed to a variety of medical conditions that range in both complexity and severity. It is postulated that these doctors are therefore at a higher risk of making medical errors than specialists.
Studies have shown that primary care doctors face uncertainty when interpreting clinical laboratory reports.
This explorative study investigated the problems and challenges that primary healthcare doctors in South Africa face with the interpretation of clinical laboratory test results. A secondary aim of the study was to identify potential countermeasures to address these challenges.
Ethical approval was granted by the Commerce Faculty Ethics in Research Committee, Graduate School of Business, University of Cape Town. Anonymity was maintained throughout the process. Informed consent was obtained through a cover letter containing the survey link. The surveys were completed online, with all responses anonymised through the system. The researchers did not have access to respondents’ identifying or personal information.
This research was based on a study conducted by Hickner et al., entitled ‘Primary care physicians’ challenges in ordering clinical laboratory tests and interpreting results’.
The original nineteen-part questionnaire was reduced to nine sections to focus on the challenges that primary healthcare doctors face with clinical laboratory result interpretation. The survey themes included doctors’ uncertainty with interpreting test results, mechanisms they use to overcome this uncertainty, challenges with appropriate result interpretation, and perceived solutions to interpreting test results. The questionnaire categories were as follows: (1) demographic information; (2) information about the doctor’s practice; (3) interpretation uncertainty; (4) the diagnostic evaluation process; (5) laboratory consultation; and (6) test utilisation enablers. Questions related to the doctor’s practice included: whether the doctor was a general practitioner or specialist; the number of years in practice; the predominant categories of tests ordered (i.e. diagnostic tests, chronic disease monitoring, or routine screening); the number of patients seen per week; the number of tests ordered per week; and the number of tests per week that were associated with interpretation uncertainty. Two questions were added to the demographic section of the questionnaire to determine whether the South African doctor worked in a rural, semi-urban or urban practice and whether he/she predominantly made use of private pathology laboratories or the parastatal (National Health Laboratory Service) laboratory. However, it was not ascertained whether the majority of patients seen were hospitalised or out-patients. Responses were predominantly chosen from a list of five-point graded options, but there was space given for open-ended responses. Response options ranged from: ‘extremely useful’ to ‘not at all useful’, ‘extremely important’ to ‘not at all important’, ‘extremely well’ to ‘not at all well’, and ‘extremely problematic’ to ‘not at all problematic’.
For this cross-sectional study, questionnaires were sent out electronically using a survey link to the approximately 7800 primary care doctors in the South African Medical Association database. The survey was sent on 13 October 2015 and remained open for responses until 13 November 2015.
Response data from the surveys were exported to Microsoft Excel (Microsoft Corp., Redmond, Washington, United States) and analysed using the IBM SPSS Statistics® package (IBM SPSS Statistics for Macintosh, Version 22.0.; IBM Corp., Armonk, New York, United States). The qualitative responses were analysed quantitatively using descriptive statistics to determine relative frequencies. Results presented were based on the number of respondents who selected the top two responses from the five-point scale – namely, ‘extremely and very useful’, ‘extremely and very important’, ‘extremely and very well’, or ‘extremely and very problematic’. Open-ended responses received were reported as ‘other’ in the figures below.
Of the approximately 7800 questionnaires sent out, 552 completed questionnaires were received, equating to a response rate of 7%. Incomplete questionnaires were excluded from the analysis, so as not to skew the results. Although the survey was sent to doctors registered in the South African Medical Association database as general and/or independent practitioners, this database included a few doctors who were either in training or were qualified specialists.
Respondents’ practice characteristics and laboratory test utilisation information, South Africa, 13 October 2015–13 November 2015.
Variables | Number of respondents | Result | Range |
---|---|---|---|
Average number of years in practice | 552 | 13 years | 1–47 years |
General or specialist practitioner | |||
General Practitioner | 513 | 93% | - |
Specialist |
39 | 7% | - |
Location of practice | |||
Urban | 298 | 54% | - |
Peri-urban | 133 | 24% | - |
Rural | 121 | 22% | - |
Type of laboratory utilised most frequently | |||
Government/Parastatal (National Health Laboratory Service) | 304 | 55% | - |
Private | 248 | 45% | - |
Predominant categories (50–100% of tests) ordered by individual doctors | |||
Diagnostic tests | 248 | 45% | - |
Chronic disease monitoring | 138 | 25% | - |
Routine screening | 94 | 17% | - |
No predominance | 72 | 13% | - |
Average number of patients seen per week | 552 | 115 patients | 3–500 patients |
Average number of laboratory tests ordered per week | 552 | 24 tests | 0–200 tests |
Average number of tests per week for which there is uncertainty in result interpretation | 552 | 4 tests | 0–30 tests |
Fields of specialty included: family medicine, internal medicine, emergency medicine, palliative medicine, HIV care, sports medicine, psychiatry, anaesthetics, paediatrics, public health, general surgery, and neurosurgery.
The challenges that doctors experienced with laboratory test results (
Challenges that doctors face when using laboratory test results, South Africa, 13 October 2015–13 November 2015.
Although not strictly related to challenges with laboratory test results, a few respondents did report challenges with laboratory access and financing in the open-ended part of this section. These challenges included: difficulties for rural hospitals or practices to get samples to a laboratory, the unavailability of specialised tests (e.g., B-type natriuretic peptide and
The majority of doctors (66%) typically used a core set of 20 or fewer clinical laboratory diagnostic tests. When faced with diagnostic uncertainty, most respondents (67%) reported always double-checking with another doctor or electronic resources (e.g., UpToDate, WebMD, patient.co.uk, etc.) if they doubted their decision. Even when confident in their pre-test diagnoses, 42% of doctors would still think ‘what else could it be?’. While 58% of clinicians were concerned about over-testing their patients, only 33% were concerned about under-testing patients.
When faced with diagnostic uncertainty in a difficult or unusual case (
Tactics employed by doctors to deal with test interpretation uncertainty, South Africa, 13 October 2015–13 November 2015.
In a variety of contexts, most respondents found communication with the laboratory to be useful (
Usefulness of laboratory communication/consultation, South Africa, 13 October 2015–13 November 2015.
In general, only about one-third of respondents noted very important reasons as to why they did not frequently contact laboratory professionals. These reasons included: difficulties in contacting the person who could answer their questions (36%); not knowing whom to contact (33%); difficulties in getting through to the laboratory (32%); and not having a working relationship with laboratory professionals (27%). Only 11% reported that they did not contact the laboratory because they felt that they had received unreliable information during previous interactions. A small number (2%) of individuals reported specific problems with public sector laboratories wherein they felt that laboratory staff were unhelpful regarding lost or rejected specimens, and inaccurate, delayed or urgent results. A lack of access to pathologists at certain regional laboratories was also noted as a problem.
Test utilisation enablers (
Usefulness and availability of test utilisation enablers, South Africa, 13 October 2015–13 November 2015.
When compared to urban and peri-urban respondents, rural doctors experienced considerably lower availability of interpretive comments, information on test performance characteristics, trending of laboratory results (when previous results are compared with current results), and reflex testing (a test performed by the laboratory in response to results from a previous test) (
Availability of test utilisation enablers for the urban, peri-urban and rural doctor cohorts,
Test Utilisation Enablers | Average ± Standard Deviation |
Urban Doctors |
Peri-urban Doctors |
Rural Doctors |
||||
---|---|---|---|---|---|---|---|---|
% | % | (%) | (%) | |||||
Interpretive comments | 414 ± 39 | 75 ± 7 | 238 | 80 | 102 | 77 | 79 | 65 |
Specialty organisation/society guidelines | 359 ± 11 | 65 ± 2 | 194 | 65 | 84 | 63 | 81 | 67 |
Dedicated laboratory phone line for questions | 215 ± 50 | 39 ± 9 | 110 | 37 | 49 | 37 | 63 | 52 |
Local practice/institution clinical algorithms | 364 ± 39 | 66 ± 7 | 200 | 67 | 97 | 73 | 71 | 59 |
Information on test performance characteristics | 177 ± 33 | 32 ± 6 | 98 | 33 | 49 | 37 | 31 | 26 |
Trending of laboratory results | 160 ± 28 | 29 ± 5 | 95 | 32 | 40 | 30 | 27 | 22 |
Reflex testing | 182 ± 28 | 33 ± 5 | 107 | 36 | 44 | 33 | 32 | 26 |
The cohorts consisted of 298 (54%) urban, 133 (24%) peri-urban and 121 (22%) rural medical practices.
Open-ended feedback on result interpretation was also elicited from respondents. Results included acknowledgement that the ‘interpretation of results are [sic] critical to reliably apply the blood results to our patients in terms of diagnosis, screening and monitoring their pathology’. A number of respondents felt that the interpretive comments currently received were not specific to patients’ age, sex, clinical picture or previous results, but were instead based on general information. Furthermore, these comments did not include recommendations for further testing or treatment. Specific challenges were noted in the interpretation of microbiology, serology (especially hepatitis B), and discordant HIV results. In contrast to the challenges noted, certain private laboratories were identified and commended for the inclusion (where necessary) of interpretive comments written by pathologists.
This study found that primary care doctors in South Africa experience challenges with laboratory test result interpretation in approximately 17% of their diagnostic encounters. By comparison, a similar study conducted in the United States found that primary care physicians reported uncertainty in 8.3% of diagnostic encounters. This emphasises the need for improved mechanisms and countermeasures to aid South African doctors with result interpretation.
The most common general challenges with laboratory test results reported by primary care doctors are related to receiving and accessing results – namely, not receiving results timeously and previous results not being easily available. Literature shows that over 80% of laboratories receive complaints about turn-around times,
Although this study focused on the post-analytical phase of laboratory testing, the survey raised two questions around the analytic testing process and whether the clinicians experienced challenges with ‘suspected errors in laboratory results’ and ‘results inconsistent with the patient’s symptoms and history’. Inconsistent results were reported as a challenge by 27% of respondents and possible laboratory errors by 28%. However, this did not include inquiry into errors occurring in the pre-analytical phase of testing, such as the mislabeling of samples. It has been found that pre-analytical errors account for 55% of laboratory errors causing a missed or delayed diagnosis.
Our study found that primary care doctors find consultation with other clinicians or laboratory professionals to be an important mechanism in aiding test result interpretation. The majority of respondents reported a dedicated laboratory phone line to be an important test utilisation enabler, and, although this was only available to less than 40% of the total study population, over half of the rural doctor cohort had access to this service. This suggests that laboratories based in rural areas are trying to leverage their limited resources. A review of literature found that failures in communication between clinicians and the laboratory could negatively impact patient safety.
The majority of survey respondents reported interpretive comments to be the most useful test utilisation enabler. Interpretive comments are added to a laboratory report in order to provide further information on the result and to aid in the diagnostic process. These comments can be provided by a qualified pathologist through technology-based interpretive algorithms and expert systems or through the addition of a ‘canned’ comment. A ‘canned’ comment is pre-written text that is added onto all results for a specific test, regardless of the actual result or the patient’s clinical history, and is considered to be the least useful form of interpretive commenting.
The majority of respondents also reported guidelines or algorithms to be useful test-utilisation enablers. Clinical algorithms and practice guidelines are developed to provide a standardised, evidence-based approach to clinical processes in order to reduce error rates, improve clinical effectiveness, and enhance the quality of patient care.
It has been found that clinicians are often unaware of whether their diagnoses at the time when they are making them are correct or erroneous.
Response rates to surveys are reported to be 10%–20%,
Primary care doctors in South Africa acknowledge that they experience uncertainty when interpreting certain clinical laboratory test results. The most useful countermeasures and mechanisms identified by the doctors to improve this included: the addition of patient-specific interpretive comments; the availability of national or international guidelines or local clinical algorithms; and enhanced communication with the laboratory through a dedicated phone line. The ultimate benefit of enhanced test result interpretation would be reduced diagnostic error rates and a more efficient and effective primary healthcare system, which would reduce the rates of referral for secondary and tertiary levels of care.
We would like to thank all the respondents who took the time to complete this survey. We also wish to thank Ms Marilyn Myburgh from the South African Medical Association for her assistance with distributing the survey and Dr Julie Taylor of the US Centers for Disease Control and Prevention for allowing us use of the questionnaire.
The authors declare that they have no financial or personal relationships which may have inappropriately influenced them in writing this article.
None.
N.V. was the project leader who co-designed the project, collected and analysed the data, and contributed to writing and preparing the manuscript. N.H.B.F. was the academic supervisor who co-designed the project and contributed to preparing the manuscript.