A multi-method study of factors associated with hospital information system success in South Africa.

A combination of interpretivist and positivist techniques was used to develop and refine a conceptual model of factors associated with computerised hospital information system (CHIS) success in South Africa. Data from three case studies of CHIS use in level 2 public sector hospitals were combined to develop a conceptual model containing seven factors associated with CHIS success at hospital level. This conceptual model formed the basis of a fourth case study which aimed to confirm and refine the initial conceptual model. In the third phase of the study, a survey of CHIS use was conducted in 30 hospitals across two South African provinces, each using one of three different CHISs. Relationships between hospital-level factors of the conceptual model and user assessment of CHIS success were examined. A revised conceptual model of CHIS use was developed on the basis of the survey results. The use of a multi-method approach made it possible to generalise results from the case studies to multiple CHIS implementations in two provinces.

Factors associated with health information system success: results of a survey of hospitals in South Africa.

A survey of computerised hospital information system (CHIS) use was conducted in two South African provinces, in order to test a conceptual model of CHIS use developed in previous phases of this study. Relationships between factors of the conceptual model and user assessment of CHIS success; and between pairs of conceptual model factors, were derived from the survey data. The results confirmed that factors of the conceptual model were associated with CHIS success. Analysis of the relationships between factors yielded results which supported some of the conceptual model relationships, and were inconclusive for others. None of the conceptual model relationships was contradicted by the survey results. Further investigation is required to demonstrate statistical relationships between factors of the conceptual model more conclusively. The results to date support arguments for the applicability of the conceptual model of CHIS use beyond the study hospitals to other level 1 and level 2 hospitals in South Africa.

A conceptual model of computerised hospital information system (CHIS) use in South Africa.

The aim of this project is to contribute to limiting the risk of CHIS (computerised hospital information system) failure by identifying factors which are associated with the successful implementation of CHISs in district and regional hospitals in South Africa (SA). Case studies were conducted in four regional hospitals in the Western Cape province of SA to obtain data about user perceptions of the success or lack of success of the CHISs in use. A conceptual model of CHIS use has been developed based on the results of the case studies, in order to assist in the interpretation of the differing experiences across the hospitals. Key factors in the conceptual model are perception of usefulness of the CHIS and management commitment to ensuring CHIS success, which in turn are related to effective use of CHIS and/or CHIS outputs, and allocation of resources for CHIS further development. Further development of the model will be influenced by the next phase of this project: a survey of district and regional hospitals in two SA provinces.

Modelling the effect of limited or vulnerable resources on the use of computerised hospital information systems (CHISs) in South Africa.

A conceptual model of computerised hospital information system (CHIS) use has been developed, based on the results of case studies in four South African regional (level 2) hospitals, interviews with local experts, and related work on modelling and evaluation of health information systems. In addition to factors within hospitals, factors which reflect the decisionmaking and resource allocation processes at provincial level have also been included in the model. The applicability of the model is demonstrated through an analysis of the effects of limited or vulnerable resources on CHIS implementation and use at hospital level. Some potential approaches to overcoming these effects are suggested.

Adult tonsillectomy--are long waiting lists putting patients at risk?

There is a paucity of data on morbidity associated with long waiting lists for adult tonsillectomy. The aim of this study was to assess the morbidity associated with long waiting lists for adult tonsillectomy in a developing world setting. Of 350 patients on the waiting list at Groote Schuur Hospital for 18 months or more, only 55 were contactable. This low yield (15.7%) from the telephonic survey highlighted the difficulty of managing long waiting lists efficiently in a developing world setting. As only 1/55 patients on the waiting list had a complication (quinsy), it appears to be safe to delay tonsillectomy in adult patients. Only half of patients ultimately required tonsillectomy because of a natural reduction in the number of episodes of tonsillitis with time. In order to avoid unnecessary tonsillectomy we need to find better prognosticators to identify the subgroup of adult patients likely to have continued recurrent tonsillitis.

Effect of losartan versus candesartan on uric acid, renal function, and fibrinogen in patients with hypertension and hyperuricemia associated with diuretics.

BACKGROUND: Hyperuricemia may counter benefits of blood pressure (BP) reduction, although this is controversial. METHODS: We examined the effects of candesartan and losartan on uric acid, creatinine, and fibrinogen. Patients with hypertension and serum uric acid > or = 0.42 mmol/L (7 mg/dL) associated with diuretics were randomized to receive losartan 50 to 100 mg or candesartan 8 to 16 mg for 24 weeks. At randomization and after 24 weeks, systolic and diastolic BP, serum uric acid, creatinine, and fibrinogen were measured. RESULTS: A total of 59 patients were entered into the study (30 in the losartan and 29 in the candesartan group). Mean systolic and diastolic BP were reduced in the candesartan group, from 156 mm Hg at baseline to 132 mm Hg at 24 weeks, and from 90.9 to 80.8 mm Hg respectively, P < .0001), and in the losartan group from 150.3 to 132 mm Hg and from 89.6 to 77.6 respectively, P < 0001). Overall mean values of fibrinogen levels were again reduced from 4.39 g/L at baseline to 4.01 g/L at 24 weeks (P < .02). Mean values of serum uric acid in the losartan and candesartan groups were similar at baseline (0.44 and 0.46 mmol/L, respectively), but they were lower in the losartan group after 24 weeks (0.39 and 0.48 mmol/L, P = .01). Twelve patients (44%) in the candesartan group had a 10% increase in serum creatinine compared with four patients (14.2%) in the losartan group (P < .02). CONCLUSIONS: Candesartan and losartan lowered BP, but only losartan reduced uric acid. The lowering of fibrinogen in both groups may explain the reduction in stroke with angiotensin receptor blockers. The effect of persistent hyperuricemia on renal function requires further study.

Amiodarone-induced thyroid dysfunction.

BACKGROUND: Little is known about the frequency of thyroid dysfunction (TD) associated with amiodarone therapy in southern Africa. OBJECTIVES: To determine the incidence of TD in a cohort of patients initiated on amiodarone therapy at a cardiac clinic in Cape Town, South Africa, believed to be an iodine-replete area. PATIENTS: Pharmacy records were used to obtain the names of patients who received amiodarone between November 1999 and December 2002. RESULTS: The sample size was 194, but data analysis was limited to the 163 patients for whom there were complete data. The mean age +/- standard deviation (SD) was 59.0 +/- 15.0 years (range 22 - 89 years). There were 67 female and 96 male patients. The indications for amiodarone therapy were supraventricular tachycardias (N = 102, 62.6%), ventricular tachycardia (N = 55, 33.7%), and prophylaxis against tachycardias (N = 3, 1.8%). The indication was uncertain in 3 patients (1.8%). The median duration of amiodarone treatment was 679.0 days (quartile deviation (QD) 1172 days, range 3 - 6425 days) in the whole cohort. The median duration of amiodarone therapy until new TD was 943 days (QD 1185 days), significantly longer than in patients who remained euthyroid (547 days, QD 1135 days) (P = 0.05). There were 45 new TD cases (27.6%): 11 patients (6.7%) were thyrotoxic, 1 (0.6%) transient thyrotoxicosis, 1 (0.6%) subclinical hyperthyroidism, 13 (8.0%) had subclinical hypothyroidism, 12 (7.4%) hypothyroidism and 7 (4.3%) had minor changes in thyroid function. CONCLUSIONS: We found a high incidence of new-onset TD, similar to the highest rates reported internationally. Local factors responsible for this need to be investigated.

Prolonged gastric emptying half-time and gastric hypomotility after drug overdose.

A prospective study was undertaken to determine if gastric motility and emptying are altered by the ingestion of overdoses of tricyclic antidepressants, acetaminophen, opioid-acetaminophen mixtures, carbamazepine or phenytoin. Gastric scintigraphy was used to measure gastric emptying half-time and assess gastric motility in 104 patients at initial presentation and again at follow-up (n = 85). Patients were imaged for 5 hours after being given 20 MBq of 99mTc tin colloid to drink. Drug serum levels were measured on all patients at initial presentation and at follow-up. We observed markedly prolonged gastric emptying half-times and severe hypomotility at initial presentation compared with follow-up in the vast majority of patients, except for a small group of patients with phenytoin poisoning. Twelve patients had gastric emptying half-times of over 300 minutes, a further 14 had half-times of over 200 minutes and 21 others had half-times of over 120 minutes. Poisoning is associated with hypomotility and a marked delay in gastric emptying that could influence the clinical course and patient management. These abnormalities may not be due to a direct effect of the ingested drug and factors such as stress may play a role.

Overt hypoadrenalism is uncommon in patients with stage 3 and 4 bronchogenic carcinoma.

INTRODUCTION: Lung cancer is the leading cause of cancer mortality in most countries. The adrenal glands are common sites of metastatic lung cancer as approximately 40% of subjects with stage 4 bronchogenic carcinoma have adrenal metastases. The prevalence of biochemical hypoadrenalism is, however, remarkably poorly documented. OBJECTIVES: Our study aimed to determine the prevalence of primary hypoadrenalism, as defined by a subnormal cortisol response to the 250 micrograms adrenocorticotrophic hormone (ACTH) stimulation test, in patients with stage 3 and 4 lung cancer. METHODS: Thirty patients with stage 3 and 4 bronchogenic carcinoma were prospectively recruited from the bronchus clinic. Demographic data and electrolytes were recorded and each patient had a 250 micrograms ACTH stimulation test to determine the prevalence of overt adrenal insufficiency, defined as a +30 minute cortisol of less than 550 nmol/l. RESULTS: The median age and quartile deviation was 62 (10) years and the median basal cortisol was 429.5 (321) nmol/l. The median peak cortisol was 828.5 (342) nmol/l (range 536-1,675 nmol/l). Twenty-eight patients (93.3%) had an appropriate rise of cortisol to greater than 550 nmol/l following 250 micrograms ACTH stimulation. Two patients (6.7%) had mild primary adrenal failure with a peak cortisol between 500 and 550 nmol/l associated with a raised plasma ACTH concentration (131.4 and 10.5 pmol/l, normal 2.2-10 pmol/l). Twenty-eight patients (92.9%) were normonatraemic, while the two hyponatraemic patients had biochemical evidence of the syndrome of inappropriate antidiuretic hormone secretion. CONCLUSION: In conclusion, despite evidence that the adrenal glands of patients with disseminated bronchogenic carcinoma are frequently affected by metastatic disease, biochemical evidence of clinically significant hypoadrenalism is relatively uncommon and is not accurately predicted by electrolyte abnormalities.

Pre- and intra-operative staging of the neck in a developing world practice.

Most patients with squamous cancer of the head and neck treated at Groote Schuur Hospital are from a poor socio-economic background with a high prevalence of tuberculosis (TB), human immunodeficiency virus (HIV) and other infections that may cause cervical lymphadenopathy resulting in overstaging of the neck. A retrospective review of 186 patients requiring therapeutic and elective neck dissection was undertaken and the sensitivity and specificity of clinical and intra-operative staging of the neck determined. Results showed overall sensitivity of staging at 80.1 per cent. Specificity was 52.2 per cent. Staging of the N(1), N(2b) and N(2c) necks had positive predictive values of 53.2, 65.8 and 68.2 per cent respectively. Occult nodal metastases were present in 32 per cent elective neck dissections (END)s. Specificity of intra-operative staging of the N(0) neck was 33.3 per cent and sensitivity was 72.4 per cent. Conclusions were that our indications for elective neck dissection are appropriate. The high false positive rates for staging of the N(1), N(2b) and N(2c) necks, necessitate a change in management strategy.

Tracheoesophageal speech in a developing world community.

OBJECTIVES: To determine the tracheoesophageal speech results in a Third World medical practice; to examine the impact of socioeconomic status, literacy, and proximity to specialist services on tracheoesophageal speech; to assess whether these factors should affect patient selection for fistula speech; and to determine guidelines for voice prosthesis selection. DESIGN: Retrospective analysis. SETTING: Groote Schuur Hospital, Cape Town, South Africa, which serves a Third World community. PATIENTS: Ninety-seven consecutive patients who underwent total laryngectomy between January 1, 1996, and October 1, 1998. Patients who undergo total laryngectomy routinely have a primary tracheoesophageal fistula created for speech. MAIN OUTCOME MEASURES: Speech outcomes after total laryngectomy; tracheoesophageal speech in relation to social class, literacy, and proximity to specialist services; and experience with removable and indwelling valves. RESULTS: Fifty-nine (81%) of 73 patients acquired useful speech. Speech outcome was not affected by employment status or proximity to specialist services. Although speech was affected by literacy and housing, several illiterate shack dwellers acquired good speech. Average device life of removable prostheses was 16 weeks (>4 months in 35% [64/183]). Indwelling prostheses had an average life of 28 weeks. CONCLUSIONS: Tracheoesophageal speech results in a Third World community equate with those in the Developed World. All patients who undergo laryngectomy and have adequate manual dexterity and cognitive function should be given a trial of fistula speech. Removable voice prostheses can successfully be used as indwelling prostheses.