Prevalence of Risk of Malnutrition in Hospitalised Adult Patients in a Tertiary Hospital Setting in South Africa

Moens, Merel (2016-03)

Thesis (MNutr)--Stellenbosch University, 2016.

Thesis

ENGLISH SUMMARY : Introduction: Hospital malnutrition was first identified by Charles Butterworth in 1974, referring to malnutrition often being overlooked, underdiagnosed and consequently undertreated. This is still a current problem, with worldwide prevalence of malnutrition ranging from 15–76% among adults. Hospital malnutrition is associated with increased cost of care, complications, increased length of stay, mortality and poor quality of life compared to well-nourished patients. South Africa’s hospitalised population is at an increased risk of malnutrition, due to high poverty levels and the quadruple burden of disease. The aim of this study was to determine the prevalence of risk of malnutrition in hospitalised patients in a South African Hospital setting. Methods: Patients that were admitted (≤48 hours) and (≥18 years old) were eligible for inclusion. The prevalence of risk of malnutrition was assessed using three different screening tools (NRS-2002, SGA and AMDT) on admission and discharge (if hospitalised ≥7days). The prevalence of risk of malnutrition, related outcomes and the number of referrals for nutritional support were documented. The included wards were assessed for availability of nutrition protocols and resources needed to implement nutrition intervention using an observational checklist. Results: On admission, a total of 403 patients were included (males 52.9%). The mean age was 45.5 years ±16.6SD. There was an even distribution between patients from surgical (n=192) and medical wards (n=190), with gynaecology (n=21) contributing a small number of patients. The prevalence of risk of malnutrition on admission ranged depending on the screening tool used: NRS-2002 (59.1%; n=237), AMDT (62.9%; n=252) and SGA (56.6%; n=228). The mean length of stay was 6.9 days ±5.9SD, with a significant difference (p<0.01) in length of stay between malnourished patients (mean 7.4 days ±6.1SD) and well-nourished patients (5.2 days ±4.8SD). On discharge, 92 patients were included (males 52.8%). Most patients (64%; n=59) endured a complication, with significantly more complications (p=0.048) among the malnourished (mean 1.7 ±1.6SD) when compared to the well-nourished (mean 0.8±1.3SD).Patients ‘at risk’ were diagnosed with infectious and gastrointestinal diseases, cancer, or had abdominal surgery, making these high-risk disease categories for malnutrition. The prevalence of risk of malnutrition was higher within the discharge sample, regardless of which tool was used: NRS-2002 (73.8%; n=62), SGA (65.2%; n=60) and AMDT (79.3%;n=73). Despite the high prevalence of malnutrition, the nutrition referrals were poor, with only 1.3% (n=5) being referred on admission, and 9.8% (n=9) on discharge. The AMDT was the only tool that had good validity (sensitivity 83.9%, specificity 80.2%) and interrater agreement (k=0.62) when using the SGA as reference. Similarly, the NRS-2002 had fair validity (sensitivity 73.8% and specificity 51.8%) but poor inter-rater agreement (k=0.24). Lastly, the hospital setting had a poor nutrition-care environment as none of the wards (n=28) had nutrition protocols, nor screening tools available at ward level. Scales were available (96.4%; n=27), but 22.2% (n=6) were not in working condition. Stadiometers were not readily available (42.9%; n=12). The mean number of patients per ward was 43 ±17.7SD, with only an average of 11 ±2.5SD nurses on duty per ward, indicating a shortage of nurses for adequate patient care. Conclusion: The prevalence of nutritional risk and malnutrition is very high in the hospital setting, regardless of screening tool used, and is associated with unfavourable patient outcomes.

AFRIKAANSE OPSOMMING : Inleiding: Hospitaal wanvoeding was onder aandag gebring deur Charles Butterworth in 1974. Hy het verwys na wanvoeding wat gereeld misgekyk, nie gediagnoseer en nie behandel word nie. Dit is steeds die geval vandag met wereld wye prevalensies van wanvoeding in die omgewing van 15-76% onder volwassenes. Hospitaal wanvoeding word geassosieer met verhoogde behandelingskoste a.g.v. die mediese intervensies benodig, komplikasies verlengde lengte van hospitalisasie, mortaliteit en swak kwaliteit van lewe in vergelyking met goed-gevoede ewekniee. Die hospitaal populasie in Suid-Afrika het ‘n verhoogde risiko om wanvoeding te ontwikkel, a.g.v. die hoe voorkoms van armoede en viervuldige siektelas. Die doel van die studie was om die prevalensie van risiko vir wanvoeding in gehospitaliseerde pasiente in Suid-Afrika te bepaal. Metodes: Pasiente wat toegelaat is binne die afgelope 48 uur en ≥18 jaar in ouderdom was geskik vir insluiting. Die prevalensie van risiko vir wanvoeding is bepaal deur drie verskillende siftingshulpmiddels (NRS-2002, SGA en AMDT) met toelating en by ontslag (indien hospitalisasie ≥7dae). Die prevalensie van risiko vir wanvoeding, verwante uitkomste (komplikasies, lengte van hospitalisasie) en die aantal verwysings vir voedingondersteuning is deurlopend aangeteken. Alle sale is evalueer vir beskikbaarheid van voedingprotokolle en hulpmiddele wat gebruik word om voedingondersteuning te implimenteer d.m.v. ‘n kontrolelys. Resultate: ‘n Totaal van 403 pasiente is ingesluit met toelating (mans 52.9%). Die gemiddelde ouderdom was 45.5 ±16.6SD. Daar was ‘n gelyke verspreiding tussen pasiente van chirurgiese (n=192) en mediese sale (n=190), met ‘n kleiner bydrae van ginekologie (n=21). Die prevalensie van wanvoeding het gewissel afhangend van die siftingshulpmiddels gebruik; NRS-2002 (59.1%; n=237), AMDT (62.9%; n=252) en SGA (56.6%; n=228). Die gemiddelde lengte van hospitalisasie was 6.9dae ±5.9SD, met ‘n beduidende verskil (p<0.01) in lengte van hospitalisasie tussen wangevoede (gemiddel 7.4 dae ±6.1SD) en goed-gevoede pasiente (5.2 dae ±4.8SD). Met ontslag is 92 pasiente ingesluit (mans 52.8%). Die meerderheid pasiente (64%; n=59) het ‘n komplikasie ontwikkel. Wangevoede pasiente met ontslag het beduidend (p=0.048) meer komplikasies gehad (gemiddel 1.7 ±1.6SD) teenoor goed-gevoede pasiente (gemiddel 0.8 ±1.3SD). Hoe-risiko siektetoestande geassosieerd met wanvoeding in hierdie studie was infektiewe en gastrointestinale siektes, kanker en abdominale chirurgie. ‘n Hoer prevalensie vir wanvoeding risiko is gevind met die ontslag-steekproef, ongegag die hulpmiddel gebruik; NRS-2002 (73.8%;n=62), SGA (65.2%;n=60) en AMDT (79.3%;n=73). Ondanks die hoe prevalensie van wanvoeding, was die voedingverwysings swak met slegs 1.3% (n=5) pasiente wat verwys is met toelating en 9.8% (n=9) met ontslag. Wanneer die siftingshulpmiddels teenoor mekaar evalueer word, was die AMDT die enigste hulpmiddel met goeie geldigheid (sensitiwiteit 83.9%, spesifisiteit 80.2%) en tussen-hulpmiddel ooreenstemming (k=0.62) teenoor die SGA as verwysing. Die NRS-2002 het ‘n matige geldigheid getoon (sensitiwiteit 73.8%, spesifisiteit 51.8%) met swak tussen-hulpmiddel ooreenstemming (k=0.24). Laastens het die hospitaal ‘n swak voedingsorg omgewing gehad deurdat geen van die sale (n=28) voedingprotokolle in plek gehad het nie, asook geen sigtingshulpmiddels op saalvlak. Skale was teenwoordig (96.4%;n=27), waarvan 22.2% (n=6) nie in werkende toestand was nie. Lengtemeters was nie geredelik beskikbaar nie (42.9%; n=12). Die gemiddelde aantal pasiente per saal was 43 ±17.7SD, terwyl daar slegs ‘n gemiddeld van 11±2.5SD verpleegkundiges aan diens was per saal. Dit dui op ‘n verplegingtekort om voldoende pasientsorg te kan lewer. Gevolgtrekking: Die prevalensie van risiko tot en wanvoeding is baie hoog in die hospitaalomgewing ongeag die siftingshulpmiddels wat gebruik is. Wanvoeding was assosieerd met ongunstige pasientuitkomste.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/98328
This item appears in the following collections: